Statisticians_group : Messages : 4039-4068 of 4068

Hello:

An non-linear regression model is a function (non-linear) that explain most of the variation in the interest variable taking the values of a set of variables (explicative) and the dependent variable to fit a model.

Where comes the model? the model comes from the Knowledge about an especific problem, i.e. , if i have a data-set about a population growth then i look for an appropiate model for that data.

There are some tecniques to fit the model, the most popular is the non-linear least squares. And others that are based in numerical tecniques depending of the pursuit purpose. For example i'm currently looking for a way to make the model fitted closest to the sample data using a tecnique with a function based the minimun-distance.

The adjust gives estimates for the model's parameters.

(I don't know STATA but you can use nls() at R.)

best regards.

BOGARTH HDZ CRISANTY

Licenciatura en Estadística

Universidad Autónoma Chapingo

De: Manish Negi <tutor_best@...>
Para: Statisticians_group@...
Enviado: jue, diciembre 24, 2009 3:35:43 AM
Asunto: Re: [Statisticians_group] Non-linear regression.please consider...

i do not know the stata but one thing is sure that we recognize the pattern of data and then decide which regression is suitable. also exponential can be converted to linear by taking natural logarithm

From: jeeva <jeeva_3006@yahoo. co.in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Thu, 24 December, 2009 10:09:16 AM
Subject: Re: [Statisticians_ group] Non-linear regression.please consider...

hello...

what is meant by non-linear regression and where we apply that concept because while i'm reading the predicted value is coated in that one paper it as e ^(2.345) (i.e) e to the power in predicted value means we use non-linear regresion..In case the e power value is 1 means we use liner regress depending upon the data .I did not understood the exact information. so,If some of them knows how & where to use it means please convey..Including stata commands.its my request...

please consider...

with regards

jeeva.......

--- On Wed, 23/12/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] Non-linear regression
To: Statisticians_ group@yahoogroup s.co.in
Date: Wednesday, 23 December, 2009, 2:20 PM

every time we donot have linearly related data neither we can change it to liner in that case we need nonlinear regression. youe question of power is not clearcut

From: jeeva <jeeva_3006@yahoo. . co.in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Wed, 23 December, 2009 10:07:52 AM
Subject: [Statisticians_ group] Non-linear regression

Hello.

what is meant by non-linear regression and where we apply that concept.

while i'm reading the e power value is 2.345 in predicted value means we use normally regress its in my point view..If some of them knows how & where to use it means please convey..Including stata commands.its my request...

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

¡Obtén la mejor experiencia en la web!
Descarga gratis el nuevo Internet Explorer 8
http://downloads.yahoo.com/ieak8/?l=e1

#4067

From: Manish Negi <tutor_best@...>
Date:: Thu Dec 24, 2009 9:35 am
Subject:: Re: Non-linear regression.please consider...

tutor_best

Offline

From: jeeva <jeeva_3006@...>
To: Statisticians_group@...
Sent: Thu, 24 December, 2009 10:09:16 AM
Subject: Re: [Statisticians_group] Non-linear regression.please consider...

hello...

with regards

jeeva.......

--- On Wed, 23/12/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] Non-linear regression
To: Statisticians_ group@yahoogroup s.co.in
Date: Wednesday, 23 December, 2009, 2:20 PM

every time we donot have linearly related data neither we can change it to liner in that case we need nonlinear regression. youe question of power is not clearcut

From: jeeva <jeeva_3006@yahoo.. co.in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Wed, 23 December, 2009 10:07:52 AM
Subject: [Statisticians_ group] Non-linear regression

Hello.

what is meant by non-linear regression and where we apply that concept.

while i'm reading the e power value is 2.345 in predicted value means we use normally regress its in my point view..If some of them knows how & where to use it means please convey..Including stata commands.its my request...

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4066

From: jeeva <jeeva_3006@...>
Date:: Thu Dec 24, 2009 4:39 am
Subject:: Re: Non-linear regression.please consider...

jeeva_3006

Offline

hello...

with regards

jeeva.......

--- On Wed, 23/12/09, Manish Negi <tutor_best@...> wrote:

From: Manish Negi <tutor_best@...>
Subject: Re: [Statisticians_group] Non-linear regression
To: Statisticians_group@...
Date: Wednesday, 23 December, 2009, 2:20 PM

every time we donot have linearly related data neither we can change it to liner in that case we need nonlinear regression. youe question of power is not clearcut

From: jeeva <jeeva_3006@yahoo. co.in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Wed, 23 December, 2009 10:07:52 AM
Subject: [Statisticians_ group] Non-linear regression

Hello.

what is meant by non-linear regression and where we apply that concept.

while i'm reading the e power value is 2.345 in predicted value means we use normally regress its in my point view..If some of them knows how & where to use it means please convey..Including stata commands.its my request...

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4065

From: jeeva <jeeva_3006@...>
Date:: Thu Dec 24, 2009 4:38 am
Subject:: Re: Non-linear regression.please consider...

jeeva_3006

Offline

hello...

with regards

jeeva.......

--- On Wed, 23/12/09, Manish Negi <tutor_best@...> wrote:

From: Manish Negi <tutor_best@...>
Subject: Re: [Statisticians_group] Non-linear regression
To: Statisticians_group@...
Date: Wednesday, 23 December, 2009, 2:20 PM

every time we donot have linearly related data neither we can change it to liner in that case we need nonlinear regression. youe question of power is not clearcut

From: jeeva <jeeva_3006@yahoo. co.in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Wed, 23 December, 2009 10:07:52 AM
Subject: [Statisticians_ group] Non-linear regression

Hello.

what is meant by non-linear regression and where we apply that concept.

while i'm reading the e power value is 2.345 in predicted value means we use normally regress its in my point view..If some of them knows how & where to use it means please convey..Including stata commands.its my request...

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4064

From: Jay Warner at AT_T <jay.warner3@...>
Date:: Wed Dec 23, 2009 3:19 pm
Subject:: Re: I need a key to SPSS 18

jaywarner2008

Offline

1) reset the clock to a time period _after_ your first use. Then restart your computer. You may need to disable your internet connection, too. Then launch SPSS. _Maybe_ that will work.

2) Download R and learn to use it. It is quite good, and still free.

3) SPSS, like other companies, spent considerable effort to build their software and ensure that it does what is claimed of it. That means someone has to pay enough to keep them going. Those of us without this level of resources are out of luck. Welcome to _that_ club. How R manages to do so well, I don't understand, but I'm glad it's there.

Jay

On Dec 22, 2009, at 4:58:28 PM, Tom Bratt wrote:

Hi, Madan Kundu

Yes, I try that. But when I put the clock behind the program didn't work. I need help! I don't know what more to do! I need a key for SPSS that I purchase in internet or another software to do the work!

Thomasina Bratt

Jay Warner

Principal Scientist

Warner Consulting, Inc.

4444 North Green Bay Road

Racine, WI 53404-1216

USA

Ph: 262.634.9100

Fax: 262.681.1133

email: quality@...

web: www.a2q.com

The A2Q Method (tm) -- What do you want to improve today?

"Put this guy to work so he'll leave me in peace!"

#4063

From: Manish Negi <tutor_best@...>
Date:: Wed Dec 23, 2009 8:50 am
Subject:: Re: Non-linear regression

tutor_best

Offline

every time we donot have linearly related data neither we can change it to liner in that case we need nonlinear regression. youe question of power is not clearcut

From: jeeva <jeeva_3006@...>
To: Statisticians_group@...
Sent: Wed, 23 December, 2009 10:07:52 AM
Subject: [Statisticians_group] Non-linear regression

Hello.

what is meant by non-linear regression and where we apply that concept.

while i'm reading the e power value is 2.345 in predicted value means we use normally regress its in my point view..If some of them knows how & where to use it means please convey..Including stata commands.its my request...

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4062

From: jeeva <jeeva_3006@...>
Date:: Wed Dec 23, 2009 4:37 am
Subject:: Non-linear regression

jeeva_3006

Offline

Hello.

what is meant by non-linear regression and where we apply that concept.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4061

From: Tom Bratt <tombratt_sm@...>
Date:: Tue Dec 22, 2009 10:58 pm
Subject:: I need a key to SPSS 18

tombratt_sm

Offline

Hi, Madan Kundu

Yes, I try that. But when I put the clock behind the program didn't work. I need help! I don't know what more to do! I need a key for SPSS that I purchase in internet or another software to do the work!

Thomasina Bratt

#4060

From: Madan Kundu <madan4331@...>
Date:: Tue Dec 22, 2009 5:37 pm
Subject:: Re: I need a key to SPSS 18

madan4331

Offline

Did u try to backdate your system? I mean put your system clock little behind.

--- On Tue, 22/12/09, Tom Bratt <tombratt_sm@...> wrote:

From: Tom Bratt <tombratt_sm@...>
Subject: [Statisticians_group] I need a key to SPSS 18
To: Statisticians_group@...
Date: Tuesday, 22 December, 2009, 10:33 PM

Hi,

My name is Thomasina Bratt. I'm new in this group. I'm a student. I need to do a work for Statistica I in the university. I need to do in this work a Correspondence Analysis, a Optimal Scaling and a Hierarchical clusters. I purchase SPSS 18 from internet, but my license expire in 4 days. I haven't resources to buy a license and I don't know any software to do this work. So, could anyone help me? Please!

Thank you,
Thomasina

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4059

From: Madan Kundu <madan4331@...>
Date:: Tue Dec 22, 2009 5:35 pm
Subject:: Re: query

madan4331

Offline

Z-test of proportion is an Wald test that follows normal distribution approximately. Z-test is derived from asymptotic theory of Maximum Likelihood estimates (See any standard text book on Mathematical Statistics, e.g., Hogg, McKean and Craig). As you know that this approximation holds good above certain sample size and for small sample size one should restrain from using this test.

On the other hand, Fisher test is an exact test which assumes that sum of row totals and sum of column totals are fixed. This is exact in the sense that it does not assume any anything about distribution. But when sample size is large, this exact procedure may be very tedious and even softwares may take lot of time. So in that case Chi-square is a good approximation to the Fisher's exact test.

So far I talked about Binomial case, but what about multinomial case?

I never tried to construct Wald type test for multinomial case. But one can construct Wald test for multinomial case as well, but it is more common to use Fisher's exct or Chi-square test in such case. So Fisher's exact or Chi-square test is much more general than Wald test, at least by application.

Hope this will give you some idea.

Thanks & Regards

MGK

--------------

Madan Gopal Kundu
PhD Scholar
Indiana University Purdue University Indianapolis (IUPUI)

Indianapolis, Indiana 46202, USA

Web: http://mgkundu.webs.com/

Cell: 317-657-1180 317-657-1180

--- On Tue, 22/12/09, Rakesh Saroj <coolvershabhu@...> wrote:

From: Rakesh Saroj <coolvershabhu@...>
Subject: [Statisticians_group] query
To: "sd" <Statisticians_group@...>
Date: Tuesday, 22 December, 2009, 5:18 PM

How Fisher's exact or Chi-square test can be used in comparing the
event rates among two groups over standard z test? Try to generalize?

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4058

From: Patrick Burns <patrick@...>
Date:: Tue Dec 22, 2009 5:16 pm
Subject:: Re: I need a key to SPSS 18

burnsstat

Offline

You might consider using R, then you
wouldn't have a problem with needing
to pay to use the software.


Patrick Burns
patrick@...
+44 (0)20 8525 0696
http://www.burns-stat.com
(home of "The R Inferno" and "A Guide for the Unwilling S User")

Tom Bratt wrote:
>
>
> Hi,
> My name is Thomasina Bratt. I'm new in this group. I'm a student. I need
> to do a work for Statistica I in the university. I need to do in this
> work a Correspondence Analysis, a Optimal Scaling and a Hierarchical
> clusters. I purchase SPSS 18 from internet, but my license expire in 4
> days. I haven't resources to buy a license and I don't know any software
> to do this work. So, could anyone help me? Please!
>
> Thank you,
> Thomasina
>
>
>

#4057

From: Tom Bratt <tombratt_sm@...>
Date:: Tue Dec 22, 2009 5:03 pm
Subject:: I need a key to SPSS 18

tombratt_sm

Offline

Hi,

My name is Thomasina Bratt. I'm new in this group. I'm a student. I need to do a work for Statistica I in the university. I need to do in this work a Correspondence Analysis, a Optimal Scaling and a Hierarchical clusters. I purchase SPSS 18 from internet, but my license expire in 4 days. I haven't resources to buy a license and I don't know any software to do this work. So, could anyone help me? Please!

Thank you,

Thomasina

#4056

From: ashwini mahapatra <ashwini_mahapatra@...>
Date:: Tue Dec 22, 2009 2:37 pm
Subject:: Re: query

ashwini_maha...

Offline

As per my understanding , if we have just 2 samples

a z test or a T test would suffice ( pick one based on sample size)

a chi square test would only be required if there are more than 2 samples.

Thanks

Ashwini

--- On Tue, 12/22/09, Manish Negi <tutor_best@...> wrote:

From: Manish Negi <tutor_best@...>
Subject: Re: [Statisticians_group] query
To: Statisticians_group@...
Date: Tuesday, December 22, 2009, 7:55 PM

it could be used to show that the event rates are not dependent on groups or within a group the events are independent of each other

--- On Tue, 22/12/09, Rakesh Saroj <coolvershabhu@ yahoo.com> wrote:

From: Rakesh Saroj <coolvershabhu@ yahoo.com>
Subject: [Statisticians_ group] query
To: "sd" <Statisticians_ group@yahoogroup s.co.in>
Date: Tuesday, 22 December, 2009, 11:48 AM

How Fisher's exact or Chi-square test can be used in comparing the
event rates among two groups over standard z test? Try to generalize?

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4055

From: Manish Negi <tutor_best@...>
Date:: Tue Dec 22, 2009 2:25 pm
Subject:: Re: query

tutor_best

Offline

it could be used to show that the event rates are not dependent on groups or within a group the events are independent of each other

--- On Tue, 22/12/09, Rakesh Saroj <coolvershabhu@...> wrote:

From: Rakesh Saroj <coolvershabhu@...>
Subject: [Statisticians_group] query
To: "sd" <Statisticians_group@...>
Date: Tuesday, 22 December, 2009, 11:48 AM

How Fisher's exact or Chi-square test can be used in comparing the
event rates among two groups over standard z test? Try to generalize?

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4054

From: Rakesh Saroj <coolvershabhu@...>
Date:: Tue Dec 22, 2009 11:48 am
Subject:: query

coolvershabhu

Offline

How Fisher's exact or Chi-square test can be used in comparing the
event rates among two groups over standard z test? Try to generalize?

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4053

From: Manish Negi <tutor_best@...>
Date:: Tue Dec 22, 2009 6:36 am
Subject:: Re: Standard error calculation

tutor_best

Offline

se(m*e) is caculated by first calculating variance as
V(m*e)=E(m*e)^2-[E(m*e)]^2

--- On Mon, 21/12/09, STANLEY110@... <STANLEY110@...> wrote:

From: STANLEY110@... <STANLEY110@...>
Subject: [Statisticians_group] Standard error calculation
To: Statisticians_group@...
Date: Monday, 21 December, 2009, 7:34 PM

I am trying to establish the minimum acceptable characteristic of a material when tested for release to the market such that the material meets its specification at the end of its expiration period.

The material degrades and degradation is linear with time. At time of manufacture, it is nominally 100 (varies some) and the lower limit of the specification is 90. At the end of its expiration period, it must be no less than 90. The specification is 90 to 110.

For a stable material, the only variability is analytical. We can estimate the lowest release limit value by provision for analytical variability of the average test result.

Lowest Release Limit = 90 + t*s/n^1/2

s = standard deviation of the release test method.

n = number of replicate assays used for the release test method.

t = one-sided Student’s t-value;

For the case where the material degrades on storage, the Lowest Release Limit is established with provision for degradative loss, uncertainty in degradative loss, and analytical uncertainty. For this case:

Lowest Release Limit = 90 + (m*e) + t*(s_T² + s²/n)^1/2

where m = slope of the fitted line that describes the loss in per unit time

e = expiration period.

s_T = standard error of m*e

My difficulty is calculating the standard error of m*e.

Any and all help will be appreciated.

(The above expression is not quite correct because the t-value is not the same for both terms in the parentheses. )

Regards,

Stan Alekman

Let charm be thy spear, wit thine armour.

Always keep your words soft and sweet, just in case you have to eat them

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

Reply | Forward | Messages in this Topic (2)

#4052

From: STANLEY110@...
Date:: Mon Dec 21, 2009 2:34 pm
Subject:: Standard error calculation

stanalek

Offline

s = standard deviation of the release test method.

I am trying to establish the minimum acceptable characteristic of a material when tested for release to the market such that the material meets its specification at the end of its expiration period.

The material degrades and degradation is linear with time. At time of manufacture, it is nominally 100 (varies some) and the lower limit of the specification is 90. At the end of its expiration period, it must be no less than 90. The specification is 90 to 110.

For a stable material, the only variability is analytical. We can estimate the lowest release limit value by provision for analytical variability of the average test result.

Lowest Release Limit = 90 + t*s/n^1/2

n = number of replicate assays used for the release test method.

t = one-sided Student’s t-value;

For the case where the material degrades on storage, the Lowest Release Limit is established with provision for degradative loss, uncertainty in degradative loss, and analytical uncertainty. For this case:

Lowest Release Limit = 90 + (m*e) + t*(s_T² + s²/n)^1/2

where m = slope of the fitted line that describes the loss in per unit time

e = expiration period.

s_T = standard error of m*e

My difficulty is calculating the standard error of m*e.

Any and all help will be appreciated.

(The above expression is not quite correct because the t-value is not the same for both terms in the parentheses.)

Regards,

Stan Alekman

Let charm be thy spear, wit thine armour.

Always keep your words soft and sweet, just in case you have to eat them

Reply | Forward | Messages in this Topic (2)

#4051

From: Manish Negi <tutor_best@...>
Date:: Mon Dec 21, 2009 2:39 pm
Subject:: Re: where we can use statistics.............

tutor_best

Offline

master of philosophy

--- On Sun, 20/12/09, Andrew Hartley <khahstats@...> wrote:

From: Andrew Hartley <khahstats@...>
Subject: Re: [Statisticians_group] where we can use statistics..............
To: Statisticians_group@...
Date: Sunday, 20 December, 2009, 5:51 PM

Mr/Ms Manish,

Increasing bridge safety is not the sole consideration. The company needs to balance safety against costs (in terms of environmental impact, time & money, at least). Maximizing a country's bridge safety alone would entail building & maintaining bridges so intensively that we would destroy the earth (excessive pollution & expending too many resources) & bankrupt the company or perhaps even the country.

I don't know if you are addressing that request for "insight" to me or to Mr/Ms Kishore. Do you wish to guide him/her, or me?

What is the "M..PHIL IN STATISTICS" you refer to? That sounds interesting. I am certainly open to your input & further discussion.

From: Manish Negi <tutor_best@yahoo. in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Sun, December 20, 2009 9:51:18 AM
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..

YES ITS TRUE THE THING IS THAT THE MEASURE OF RISK IS INDICATED IN THEIR EFFORTS BECAUSE NOBODY WANTS ITS COMPANY TO BE BANNED THAT'S WHY THEY EMPLOY THE STATISTICIAN TO INCREASE THE SAFETY NOT TO DECREASE.. SO ITIS INTUITIVE RATHER THAN EXPLANATORY
BYE THE WAY COULD I GET SOME INSIGHT WHAT YOU ARE DOING SO THAT I CAN GUIDE YOU MORE I AM M..PHIL IN STATISTICS

--- On Sun, 20/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Sunday, 20 December, 2009, 10:17 AM

agreed; the company also has obligations to its workers. I don't know how to incorporate those into a cost-benefit analysis, though, except informally of course. That analysis cannot pretend to constitute "the entire story."

From: Manish Negi <tutor_best@yahoo. in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Sun, December 20, 2009 2:59:52 AM
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..

please note that are social obligations are also to infrastructurs company which cannot be violated. there should also be attached the probability fo risk

--- On Sat, 19/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 11:41 PM

Mr/Ms Manish,

Those were good answers.

Mr/Ms Kishore,

Please allow me to address just 1 of your areas, infrastructure companies, but in more depth. I believe statistics aims to help us make certainty statements about unknown quantities, & also to compare various allocations of resources. Assume that an infrastructure company wanted to decide whether to switch to a less expensive design for bridges. While the new design could allow them to complete more projects in a given amount of time, the company needs, of course, to evaluate how often bridges built this way might fail to pass government inspections. Statistical methods could be used to quantify the stochastic distributions of both the time savings & the inspection-failures . Whether they move to the new design could then depend on the results of a statistical cost-benefit analysis.

--- On Sat, 12/19/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, December 19, 2009, 12:39 PM

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques.. ......

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@ yahoo..co. in> wrote:

From: kishore reddy <kishore_pingili@ yahoo.co. in>
Subject: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....

i have some questions ............ .

1] how can we implement statistics in Construction companies[Infra structure companies]

2] How can we implement Statistics in Stores related works in any organization. ........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

which statistical techniques we use in the above asked questions... ........

thanks a million in advance for giving answers for my above asked questions... . .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4050

From: Andrew Hartley <khahstats@...>
Date:: Mon Dec 21, 2009 2:11 pm
Subject:: Re: where we can use statistics............. [1 Attachment]

khahstats

Offline

Jay,

you seem to be leading towards a definition of statistics or, at least, of a way of answering questions using stats. How, though, does stats differ from, say, physics or chemistry or any other science? To me, to describe answering questions using stats, in contrast to those other sciences, we must somehow incorporate partial determinacy. By the latter, I mean the idea that (in parametric inference) unknown parameters or states of affairs partially determine observed data, or that (in predictive analysis & decision analysis) data partially determine future outcomes. Somewhat more technically, a change in what we see (observations) induces a change in the probability distribution of what we do not see (unknown parameters, future observations or future utilities).

The way I think of statistics, this discipline specializes in managing & accounting for persistent partial uncertainty (persistent in the sense that, no matter how much data we collect,

1. our conclusions remain uncertain (as in estimates) or, a fortiori,

2. we actually state our conclusions (as in probability distributions of parameters) in terms of uncertainty)

You were not, admittedly, giving a formal definition of stats; however, I suggest that, even in describing its uses, we have to draw attention to stats in particular, & not to science in general. Your thoughts?

From: Jay Warner at AT_T <jay.warner3@...>
To: Statisticians_group@...
Sent: Sun, December 20, 2009 10:55:53 PM
Subject: Re: [Statisticians_group] where we can use statistics............. [1 Attachment]

Rather than focus on specific applications of "statistics, " why not
ask how to answer questions and solve problems.

In every operation we have activities that do not come out the way we
want them to. Hence, we have problems everywhere. Is there something
we might change, that _might_ reduce the problem? If not, then there
is nothing to be done; we might as well go home. If yes, then how are
we to know the change will benefit? By

(a) formulating a question in a manner that can be answered by
measurements, data, and a (statistical) analysis of it.
(b) collecting data which can be analyzed in some way, and answer the
question
(c) a method of analyzing the data, that suits the data, and will
address the original question.

My suspicion is that you are having trouble with item (a). Welcome to
the club; that is the hardest part, IMHO.

Jay
On Dec 19, 2009, at 11:30:18 AM, kishore reddy wrote:

>
>
> Hello statisticians,
>
>
>
> Where can we use statistics in our real
> life........ ....
>
>
> i have some questions ............
>
> 1] how can we implement statistics in Construction companies[Infra
> structure companies]

PERT/CPM - project management - will tell you how long a project will
take, the bottlenecks, etc. in it. Calculating the 'critical path'
with introduced bits of random variation for each activity will reveal
a distribution of likely completion dates, and give us an estimate of
the expected completion time, with a confidence interval. Some
software does this more rapidly than I can describe it. The software
will not, however, tell you which management changes to make (more
labor in one activity, for example) to reduce the modal completion date.

Analyzing the project plan's critical path with selected variations in
major activities using a Designed Experiment approach will reveal
those management actions that can seriously reduce the total project
time, and also reveal those uncontrollable variations that managers
need to watch for most closely, so they can take effective, mitigating
action. See Santell, M. P. ; Jung, J. R., Jr.; Warner, J. C.,
"Optimization in Project Coordination Scheduling Through Application
of Taguchi Methods," Project Management Journal: XXIII(3), pg. 5-16,
Sept. 1992

> 2] How can we implement Statistics in Stores related works in any
> organization. ........

How do we justify a specific marketing/sales plan? How do we decide if
a new point-of-sale sign is effective? How do we determine
differences in customer return rates associated with different
suppliers? Or customer complaints? How shall we rate (and rank)
checkout clerks in a retail store?

> 3] how can we use statistics in Voluntary
> organization. ......... ...and social organization

The Voluntary organization does not care about the effectiveness of
goods it purchases? Same issue as in for-profit firms. How do we
know the organization is meeting its own objectives? We have to
measure something, and the measurements may not be absolutely
precise. You will need some sort of statistical method to weed
through that measurement variation, I bet.

> which statistical techniques we use in the above asked
> questions... ........

Lots of them. Depends on items (a) and (b) above. This may sound
like I am making light of your questions, but I am not. You are the
one who must ask the question. It comes from your perception of a
'problem' -- a product or output or behavior that does not act the way
you want it to. Others, such as this discussion group, can help you
refine the question to suit data you can collect and work out an
analysis that will use the data and answer the question. - this last
part, (c), has the statistical techniques.

But you cannot ask about (c) until you have (a) and (b). You will
need to adjust each of them until they all 'match' up together.

So what is _your_ problem, and what might you do to fix it?

Jay

Jay Warner
Principal Scientist
Warner Consulting, Inc.
4444 North Green Bay Road
Racine, WI 53404-1216
USA

Ph: 262.634.9100
Fax: 262.681.1133
email: quality@a2q. com
web: www.a2q.com

The A2Q Method (tm) -- What do you want to improve today?

"Put this guy to work so he'll leave me in peace!"

#4049

From: "predictiveanalyticsworld" <elise@...>
Date:: Mon Dec 21, 2009 9:21 am
Subject:: Predictive Analytics World coming February to San Francisco

predictivean...

Offline

Hi Everyone,

Just a reminder to save the date: Predictive Analytics World conference, Feb.
16-17, 2010, San Francisco.

Includes keynote speakers Kim Larsen, Director Advanced Analytics at Charles
Schwab, and Andreas S. Weigend, Ph.D., Former Chief Scientist at Amazon.com.
Case studies presented by Amazon.com, AT&T, Charles Schwab, Google, IBM, PayPal
(eBay), U.S. Army, Visa, Walmart Financial Services, various U.S. government
agencies and more.

Three pre- and post-event workshops complement the core conference program.

For details on February's event, see our recent blog entry
http://www.predictiveanalyticsworld.com/blog/?p=98

For more info, visit http://www.predictiveanalyticsworld.com, e-mail us at
registration@... or call (718) 577-1202.

thanks --Elise Johnson, Predictive Analytics World

Reply | Forward | Messages in this Topic (1)

#4048

From: Tirumala B <b.tirumala@...>
Date:: Mon Dec 21, 2009 5:07 am
Subject:: Re: where we can use statistics.............

b.tirumala

Offline

we use statistics in anywhere . without statistics we cant get final results.

mainly we are using 5 streams....

1.Economic Analysis
2.Scientific Research
3.Plannig
4.Business and Commerce
5.Forecasting and Projection

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4047

From: Jay Warner at AT_T <jay.warner3@...>
Date:: Mon Dec 21, 2009 3:55 am
Subject:: Re: where we can use statistics.............

jaywarner2008

Offline

Rather than focus on specific applications of "statistics," why not
ask how to answer questions and solve problems.

In every operation we have activities that do not come out the way we
want them to.  Hence, we have problems everywhere.  Is there something
we might change, that _might_ reduce the problem?  If not, then there
is nothing to be done; we might as well go home.  If yes, then how are
we to know the change will benefit?  By

(a) formulating a question in a manner that can be answered by
measurements, data, and a (statistical) analysis of it.
(b) collecting data which can be analyzed in some way, and answer the
question
(c) a method of analyzing the data, that suits the data, and will
address the original question.

My suspicion is that you are having trouble with item (a).  Welcome to
the club; that is the hardest part, IMHO.

Jay
On Dec 19, 2009, at 11:30:18 AM, kishore reddy wrote:

>
>
> Hello statisticians,
>
>
>
>                    Where can we use statistics in our real
> life............
>
>
> i have some questions ............
>
> 1] how can we implement statistics in Construction companies[Infra
> structure companies]

PERT/CPM - project management - will tell you how long a project will
take, the bottlenecks, etc. in it.  Calculating the 'critical path'
with introduced bits of random variation for each activity will reveal
a distribution of likely completion dates, and give us an estimate of
the expected completion time, with a confidence interval.  Some
software does this more rapidly than I can describe it.  The software
will not, however, tell you which management changes to make (more
labor in one activity, for example) to reduce the modal completion date.

Analyzing the project plan's critical path with selected variations in
major activities using a Designed Experiment approach will reveal
those management actions that can seriously reduce the total project
time, and also reveal those uncontrollable variations that managers
need to watch for most closely, so they can take effective, mitigating
action.  See Santell, M. P. ; Jung, J. R., Jr.; Warner, J. C.,
"Optimization in Project Coordination Scheduling Through Application
of Taguchi Methods," Project Management Journal: XXIII(3), pg. 5-16,
Sept. 1992

> 2] How can we implement Statistics in Stores related works in any
> organization.........

How do we justify a specific marketing/sales plan? How do we decide if
a new point-of-sale sign is effective?  How do we determine
differences in customer return rates associated with different
suppliers?  Or customer complaints?  How shall we rate (and rank)
checkout clerks in a retail store?

> 3] how can we use statistics in Voluntary
> organization.............and social organization

The Voluntary organization does not care about the effectiveness of
goods it purchases?  Same issue as in for-profit firms.  How do we
know the organization is meeting its own objectives?  We have to
measure  something, and the measurements may not be absolutely
precise.  You will need some sort of statistical method to weed
through that measurement variation, I bet.

> which statistical techniques we use in the above asked
> questions...........

Lots of them.  Depends on items (a) and (b) above.  This may sound
like I am making light of your questions, but I am not.  You are the
one who must ask the question.  It comes from your perception of a
'problem' -- a product or output or behavior that does not act the way
you want it to.  Others, such as this discussion group, can help you
refine the question to suit data you can collect and work out an
analysis that will use the data and answer the question. - this last
part, (c), has the statistical techniques.

But you cannot ask about (c) until you have (a) and (b).  You will
need to adjust each of them until they all 'match' up together.

So what is _your_ problem, and what might you do to fix it?

Jay

Jay Warner
Principal Scientist
Warner Consulting, Inc.
4444 North Green Bay Road
Racine, WI 53404-1216
USA

Ph:       262.634.9100
Fax:     262.681.1133
email:  quality@...
web:    www.a2q.com

The A2Q Method (tm) -- What do you want to improve today?

"Put this guy to work so he'll leave me in peace!"

1 of 1 Photo(s)

JayWarner_MG_7485_h.gif

#4046

From: Bogarth Hernandez <crisanty@...>
Date:: Mon Dec 21, 2009 2:40 am
Subject:: Re: Frequentist vs Bayesian Stat.

crisanty

Offline

Visit Your Group Start a new topic

Thanks for all comments. Maybe i received more than i expected, but all is well-come.

Merry X-mass for all and i expect that you have a nice week.

BOGARTH HDZ CRISANTY

Licenciatura en Estadística

Universidad Autónoma Chapingo

Reply to sender | Reply to group

Messages in this topic (14)

Recent Activity:

New Members 6

________________________________________________________________________
NOTE: The Statisticians_group is for discussing on different issues related to the subject Statistics, Statistical softwares like SAS, SPSS etc. The users are encouraged to post jobs those are exclusively for Statisticians like data analyst, business analyst etc.

Please don't post or send any SPAM!

NOTE: Please don't use this mail for any personal discussion.

________________________________________________________________________

Switch to: Text-Only, Daily Digest • Unsubscribe • Terms of Use

¡Obtén la mejor experiencia en la web!
Descarga gratis el nuevo Internet Explorer 8
http://downloads.yahoo.com/ieak8/?l=e1

Reply | Forward | Messages in this Topic (15)

#4045

From: Andrew Hartley <khahstats@...>
Date:: Sun Dec 20, 2009 5:51 pm
Subject:: Re: where we can use statistics.............

khahstats

Offline

Mr/Ms Manish,

Increasing bridge safety is not the sole consideration. The company needs to balance safety against costs (in terms of environmental impact, time & money, at least). Maximizing a country's bridge safety alone would entail building & maintaining bridges so intensively that we would destroy the earth (excessive pollution & expending too many resources) & bankrupt the company or perhaps even the country.

I don't know if you are addressing that request for "insight" to me or to Mr/Ms Kishore. Do you wish to guide him/her, or me?

What is the "M..PHIL IN STATISTICS" you refer to? That sounds interesting. I am certainly open to your input & further discussion.

From: Manish Negi <tutor_best@...>
To: Statisticians_group@...
Sent: Sun, December 20, 2009 9:51:18 AM
Subject: Re: [Statisticians_group] where we can use statistics.............

YES ITS TRUE THE THING IS THAT THE MEASURE OF RISK IS INDICATED IN THEIR EFFORTS BECAUSE NOBODY WANTS ITS COMPANY TO BE BANNED THAT'S WHY THEY EMPLOY THE STATISTICIAN TO INCREASE THE SAFETY NOT TO DECREASE.. SO ITIS INTUITIVE RATHER THAN EXPLANATORY
BYE THE WAY COULD I GET SOME INSIGHT WHAT YOU ARE DOING SO THAT I CAN GUIDE YOU MORE I AM M..PHIL IN STATISTICS

--- On Sun, 20/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Sunday, 20 December, 2009, 10:17 AM

agreed; the company also has obligations to its workers. I don't know how to incorporate those into a cost-benefit analysis, though, except informally of course. That analysis cannot pretend to constitute "the entire story."

From: Manish Negi <tutor_best@yahoo. in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Sun, December 20, 2009 2:59:52 AM
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..

please note that are social obligations are also to infrastructurs company which cannot be violated. there should also be attached the probability fo risk

--- On Sat, 19/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 11:41 PM

Mr/Ms Manish,

Those were good answers.

Mr/Ms Kishore,

Please allow me to address just 1 of your areas, infrastructure companies, but in more depth. I believe statistics aims to help us make certainty statements about unknown quantities, & also to compare various allocations of resources. Assume that an infrastructure company wanted to decide whether to switch to a less expensive design for bridges. While the new design could allow them to complete more projects in a given amount of time, the company needs, of course, to evaluate how often bridges built this way might fail to pass government inspections. Statistical methods could be used to quantify the stochastic distributions of both the time savings & the inspection-failures . Whether they move to the new design could then depend on the results of a statistical cost-benefit analysis.

--- On Sat, 12/19/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, December 19, 2009, 12:39 PM

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques.. ......

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@ yahoo..co. in> wrote:

From: kishore reddy <kishore_pingili@ yahoo.co. in>
Subject: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....

i have some questions ............ .

1] how can we implement statistics in Construction companies[Infra structure companies]

2] How can we implement Statistics in Stores related works in any organization. ........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

which statistical techniques we use in the above asked questions... ........

thanks a million in advance for giving answers for my above asked questions... . .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4044

From: Manish Negi <tutor_best@...>
Date:: Sun Dec 20, 2009 2:51 pm
Subject:: Re: where we can use statistics.............

tutor_best

Offline

From: Andrew Hartley <khahstats@...>
Subject: Re: [Statisticians_group] where we can use statistics.............
To: Statisticians_group@...
Date: Sunday, 20 December, 2009, 10:17 AM

agreed; the company also has obligations to its workers. I don't know how to incorporate those into a cost-benefit analysis, though, except informally of course. That analysis cannot pretend to constitute "the entire story."

From: Manish Negi <tutor_best@yahoo. in>
To: Statisticians_ group@yahoogroup s.co.in
Sent: Sun, December 20, 2009 2:59:52 AM
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..

please note that are social obligations are also to infrastructurs company which cannot be violated. there should also be attached the probability fo risk

--- On Sat, 19/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 11:41 PM

Mr/Ms Manish,

Those were good answers.

Mr/Ms Kishore,

Please allow me to address just 1 of your areas, infrastructure companies, but in more depth. I believe statistics aims to help us make certainty statements about unknown quantities, & also to compare various allocations of resources. Assume that an infrastructure company wanted to decide whether to switch to a less expensive design for bridges. While the new design could allow them to complete more projects in a given amount of time, the company needs, of course, to evaluate how often bridges built this way might fail to pass government inspections. Statistical methods could be used to quantify the stochastic distributions of both the time savings & the inspection-failures . Whether they move to the new design could then depend on the results of a statistical cost-benefit analysis.

--- On Sat, 12/19/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, December 19, 2009, 12:39 PM

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques.. ......

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@ yahoo..co. in> wrote:

From: kishore reddy <kishore_pingili@ yahoo.co. in>
Subject: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....

i have some questions ............ .

1] how can we implement statistics in Construction companies[Infra structure companies]

2] How can we implement Statistics in Stores related works in any organization. ........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

which statistical techniques we use in the above asked questions... ........

thanks a million in advance for giving answers for my above asked questions... . .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4043

From: Andrew Hartley <khahstats@...>
Date:: Sun Dec 20, 2009 10:17 am
Subject:: Re: where we can use statistics.............

khahstats

Offline

agreed; the company also has obligations to its workers. I don't know how to incorporate those into a cost-benefit analysis, though, except informally of course. That analysis cannot pretend to constitute "the entire story."

From: Manish Negi <tutor_best@...>
To: Statisticians_group@...
Sent: Sun, December 20, 2009 2:59:52 AM
Subject: Re: [Statisticians_group] where we can use statistics.............

please note that are social obligations are also to infrastructurs company which cannot be violated. there should also be attached the probability fo risk

--- On Sat, 19/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 11:41 PM

Mr/Ms Manish,

Those were good answers.

Mr/Ms Kishore,

Please allow me to address just 1 of your areas, infrastructure companies, but in more depth. I believe statistics aims to help us make certainty statements about unknown quantities, & also to compare various allocations of resources. Assume that an infrastructure company wanted to decide whether to switch to a less expensive design for bridges. While the new design could allow them to complete more projects in a given amount of time, the company needs, of course, to evaluate how often bridges built this way might fail to pass government inspections. Statistical methods could be used to quantify the stochastic distributions of both the time savings & the inspection-failures . Whether they move to the new design could then depend on the results of a statistical cost-benefit analysis.

--- On Sat, 12/19/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, December 19, 2009, 12:39 PM

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques.. ......

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@ yahoo..co. in> wrote:

From: kishore reddy <kishore_pingili@ yahoo.co. in>
Subject: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....

i have some questions ............ .

1] how can we implement statistics in Construction companies[Infra structure companies]

2] How can we implement Statistics in Stores related works in any organization. ........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

which statistical techniques we use in the above asked questions... ........

thanks a million in advance for giving answers for my above asked questions... . .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4042

From: Manish Negi <tutor_best@...>
Date:: Sun Dec 20, 2009 7:59 am
Subject:: Re: where we can use statistics.............

tutor_best

Offline

From: Andrew Hartley <khahstats@...>
Subject: Re: [Statisticians_group] where we can use statistics.............
To: Statisticians_group@...
Date: Saturday, 19 December, 2009, 11:41 PM

Mr/Ms Manish,
Those were good answers.

Mr/Ms Kishore,
Please allow me to address just 1 of your areas, infrastructure companies, but in more depth. I believe statistics aims to help us make certainty statements about unknown quantities, & also to compare various allocations of resources. Assume that an infrastructure company wanted to decide whether to switch to a less expensive design for bridges. While the new design could allow them to complete more projects in a given amount of time, the company needs, of course, to evaluate how often bridges built this way might fail to pass government inspections. Statistical methods could be used to quantify the stochastic distributions of both the time savings & the inspection-failures . Whether they move to the new design could then depend on the results of a statistical cost-benefit analysis.

--- On Sat, 12/19/09, Manish Negi <tutor_best@yahoo. in> wrote:

From: Manish Negi <tutor_best@yahoo. in>
Subject: Re: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, December 19, 2009, 12:39 PM

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques.. ......

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@ yahoo..co. in> wrote:

From: kishore reddy <kishore_pingili@ yahoo.co. in>
Subject: [Statisticians_ group] where we can use statistics.. .......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....

i have some questions ............ .

1] how can we implement statistics in Construction companies[Infra structure companies]

2] How can we implement Statistics in Stores related works in any organization. ........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

which statistical techniques we use in the above asked questions... ........

thanks a million in advance for giving answers for my above asked questions.... .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4041

From: Andrew Hartley <khahstats@...>
Date:: Sat Dec 19, 2009 11:59 pm
Subject:: Re: Frequentist vs Bayesian Stat.

khahstats

Offline

Mr/Ms Kishore,

Thanks for your kind words. Yes, I do consider fiducial inference to be potentially useful, & the p-values I mentioned below could be considered a tool within FI. Another example is confidence intervals. CIs, as with any frequentist method, are very difficult to interpret correctly. Any particular 95% CI is a member of a series of CIs, 95% of which contain the unknown parameter. Unfortunately, people cannot resist switching that 95% probability, which applies to the sample space, so that it refers (without justification) to the parameter space. In other words, just because 95% of the CIs contain the unknown parameter, does not mean that any particular CI contains the parameter with 95% probability. A (bayesian) highest posterior density interval (HPDI), however, does contain the parameter with 95% probability. The link between bayesianism & frequentism here is that, under some conditions, CIs are approximately the same as HPDIs.

People need to know, however, the conditions under which frequentist results can be re-interpreted in these ways. Often, people think that having a flat bayesian prior is sufficient; however, sometimes that is not true, especially when the sample space and/or parameter space are bounded or discrete.

I appreciate the reference to Jaynes, & will take a look at that book, esp. his discussion of decision theory. I found a PDF version at http://www-biba.inrialpes.fr/Jaynes/prob.html

--- On Sat, 12/19/09, Nand Kishore Singh <nk_singh1@...> wrote:

From: Nand Kishore Singh <nk_singh1@...>
Subject: Re: [Statisticians_group] Frequentist vs Bayesian Stat.
To: Statisticians_group@...
Date: Saturday, December 19, 2009, 2:46 AM

Dear Andrew!
Thanks for nice exposition to understand difference between frequentist and bayesian approach. You have mentioned “Some frequentist results, though deductive, can, under some conditions & assumptions, be interpreted inductively (bayesianly) . They are helpful in advancing science to the extent that they can be so interpreted, & to the extent that they are easier to compute than Bayesian results.” Whether you intent fiducial inference?. Can you provide some hints what are other methods and how they are easy to compute?
I will suggest book “Probabilty Therory: Logic of Science” by E.T. Jaynes (book is available in .ps file. (download software or use site to read it) . This book provide an unique insight (logical interpretation in place of frequentist) for probability and statistical inference with emphasis on ideas (in place of mathematical complexity). It also provide a rational base for priors (and methods to compute it). Jaynes had evaluated many alternative approaches (in comparison to bayesian) and found that they are not complete as they are not satisfying Cox theorem.
Nand Kishore

--- On Sat, 19/12/09, Andrew Hartley <khahstats@yahoo. com> wrote:

From: Andrew Hartley <khahstats@yahoo. com>
Subject: Re: [Statisticians_ group] Frequentist vs Bayesian Stat.
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 4:47 AM

Mr Crisanty,

Here are some of my views concerning Frequentism & Bayesianism.

Statistical inference seeks to be a means of making statements about unknown parameters, populations, & hypotheses, given data. Viz., this inference should be inductive (reasoning from the specific to the general). For example, given the particular results of the experiment we just performed, we want to say something about the general population from which we sampled our experimental units.

Frequentist results, in contrast, are statements about particular data assuming those generalities. Viz., these results are deductive (reasoning from the general to the specific). For instance, a p-value is a statement about what future data might be observed that is “more extreme” than the currently observed data, assuming the tested hypothesis. The question frequentists need to answer, then, in defending their frequentism, is “given a statement about the specific given the general, what can we say in turn about the general?”

Bayesian results, on the other hand, do constitute inductive inference, which is why I’m a Bayesian. Perhaps the commonest Bayesian result, for instance is the probability of an experimental hypothesis, given the data, viz., Pr(H | x).

2 difficulties people have with bayesianism—preventing their acceptance of bayesianism—are

Bayes theorem, on which most or all Bayesian results are based, requires prior (pre-experimental) probabilities about all unknown relevant quantities. Empiricist people believe that the data should be able to tell us what to believe & what to do, without the use of such prior probs.

Bayesian results tell us what to believe, given those prior probabilities, the data, & a model for those data. Post-modernists want to believe what they want to believe, & don’t want to be limited by the data.

Some frequentist results, though deductive, can, under some conditions & assumptions, be interpreted inductively (bayesianly) *. They are helpful in advancing science to the extent that they can be so interpreted, & to the extent that they are easier to compute than Bayesian results.

There are many schools that specialize in Bayesian stats, & offer MS/PhD programs. I suggest you Google something like “university Bayesian statistics phd program degree” & see what pops up. As Madan said, the important thing is to work with an advisor who knows a great deal about bayesianism, including the literature, unsolved problems, greatest needs for additional research, etc.

FYI, a paradigm which tries to compromise between frequentism & bayesianism is “evidentialism,” which you could research by googling something like “evidentialist Goodman Royall.”

I could go on with more of my views, but maybe that’s enough for now. Just write me if you want to discuss more.

*For example, a p-value can sometimes be interpreted as a post-experimental probability of the tested hypothesis.

--- On Fri, 12/18/09, Bogarth Hernandez <crisanty@yahoo. com> wrote:

From: Bogarth Hernandez <crisanty@yahoo. com>
Subject: [Statisticians_ group] Frequentist vs Bayesian Stat.
To: Statisticians_ group@yahoogroup s.co.in
Date: Friday, December 18, 2009, 2:31 PM

There are two approaches (that i know) to Statistics frequentist & Bayesian.

We live in a world where the level of application of Stat Theory is poor in general, but statisticians are always arguing about whether what approach is "the valid one" (this is occurr in Mexico too).

I'm a frequentist because i studied in a frequentist school, but i would like to learn bayesian stat. I think that the best statistician is one that is well-prepared to use both of them.

If someone can give me an advice about where i can do MSc. or Ph.D. with both approaches, i will be greatful (UK or USA).

BOGARTH HDZ CRISANTY

Licenciado en Estadística

Universidad Autónoma Chapingo

De: Andrew Hartley <khahstats@yahoo. com>
Para: Statisticians_ group@yahoogroup s.co.in
Enviado: jue, diciembre 17, 2009 7:24:28 PM
Asunto: Re: [Statisticians_ group] Re: LOF

Mr/Ms Luis,

Most LOF tests—such as yours, I assume—are frequentist, & frequentism has no concept or definition of “strength of evidence,” even though frequentist statisticians speak of evidence all the time. Different frequentists can use “evidence” differently &, therefore, talk right past one another, sometimes without even knowing it. The ensuing arguments are really quite meaningless. Frequentist A says no evidence exists, & Frequentist B says evidence exists, just because they either have different concepts of evidence or, equally as often, haven’t even defined evidence for themselves. It’s like me arguing with someone about the meaning of the word “actual,” just because I’m from USA & he is from Spain .

As I noted below, if you say no evidence exists for LOF because a hypothesis test failed to reject the hypothesis of perfect fit (large p-value), such reasoning would not be logical, even if we forget, for the moment, the weaknesses of hypothesis testing. A hypothesis test cannot “say there is no evidence” for the tested hypothesis.

Bayesians, on the other hand, can measure evidence precisely, but maybe you would want to hear about that on another occasion.

Perhaps what you are wondering about is why one LOF test rejected its own hypothesis of perfect fit, & why another LOF test failed to reject its own, different hypothesis of perfect fit?

--- On Thu, 12/17/09, huete.luis@ymail. com <huete.luis@ymail. com> wrote:

From: huete.luis@ymail. com <huete.luis@ymail. com>
Subject: [Statisticians_ group] Re: LOF
To: Statisticians_ group@yahoogroup s.co.in
Date: Thursday, December 17, 2009, 7:10 AM

I have two paired data. One is a sinewave ramping upward and the other is a wave ramping upward. I run a simple linear regression analysis with LOF test (null hypothesis: simple linear regression model is correct). LOF test says there is no evidence of LOF for sinewave ramping upward. Also, LOF test says there is evidence of LOF for wave ramping upward. I'm trying to find the explaination behind these results.

--- In Statisticians_ group@yahoogroup s.co.in, Andrew Hartley <khahstats@. ..> wrote:
>
> Mr/Ms Luis,
> okay; why do you say no evidence exists for LOF? What are you using to measure or reflect the evidence? Also, how do you know you have a sine wave ramping up? have you calculated means for your dependent variable, grouping by your (possibly categorized) explanatory variable?
>
> Are you saying no evidence exists for LOF because a hypothesis test failed to reject the hypothesis of perfect fit? Such reasoning would not be logical.
>
> --- On Wed, 12/16/09, huete.luis@. .. <huete.luis@ ...> wrote:
>
> From: huete.luis@. .. <huete.luis@ ...>
> Subject: [Statisticians_ group] Re: LOF
> To: Statisticians_ group@yahoogroup s.co.in
> Date: Wednesday, December 16, 2009, 2:39 PM
>
>
>
>
>
>
>
> Â
>
>
>
>
>
>
>
>
>
> simple linear regression
>
>
>
> --- In Statisticians_ group@yahoogroup s.co.in, Andrew Hartley <khahstats@ ..> wrote:
>
> >
>
> > lack of fit relative to what ideal curve(s)? Normal? Exponential? Logistic?
>
> >
>
> > --- On Wed, 12/16/09, huete.luis@ .. <huete.luis@ ...> wrote:
>
> >
>
> >
>
> > From: huete.luis@ .. <huete.luis@ ...>
>
> > Subject: [Statisticians_ group] LOF
>
> > To: Statisticians_ group@yahoogroup s.co.in
>
> > Date: Wednesday, December 16, 2009, 6:49 AM
>
> >
>
> >
>
> > Anyone know why there is no evidence of lack-of-fit when you have a sine wave ramping upward? There is lack of fit when you have a wave ramping upward. Thanks.
>
> >
>

¡Obtén la mejor experiencia en la web!
Descarga gratis el nuevo Internet Explorer 8
http://downloads. yahoo.com/ ieak8/?l= e1

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

Reply | Forward | Messages in this Topic (15)

#4040

From: Andrew Hartley <khahstats@...>
Date:: Sat Dec 19, 2009 11:41 pm
Subject:: Re: where we can use statistics.............

khahstats

Offline

Mr/Ms Manish,

Those were good answers.

Mr/Ms Kishore,

Please allow me to address just 1 of your areas, infrastructure companies, but in more depth. I believe statistics aims to help us make certainty statements about unknown quantities, & also to compare various allocations of resources. Assume that an infrastructure company wanted to decide whether to switch to a less expensive design for bridges. While the new design could allow them to complete more projects in a given amount of time, the company needs, of course, to evaluate how often bridges built this way might fail to pass government inspections. Statistical methods could be used to quantify the stochastic distributions of both the time savings & the inspection-failures. Whether they move to the new design could then depend on the results of a statistical cost-benefit analysis.

--- On Sat, 12/19/09, Manish Negi <tutor_best@...> wrote:

From: Manish Negi <tutor_best@...>
Subject: Re: [Statisticians_group] where we can use statistics.............
To: Statisticians_group@...
Date: Saturday, December 19, 2009, 12:39 PM

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques.. ......

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@ yahoo.co. in> wrote:

From: kishore reddy <kishore_pingili@ yahoo.co. in>
Subject: [Statisticians_ group] where we can use statistics.. ......... ..
To: Statisticians_ group@yahoogroup s.co.in
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....

i have some questions ............ .

1] how can we implement statistics in Construction companies[Infra structure companies]

2] How can we implement Statistics in Stores related works in any organization. ........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

which statistical techniques we use in the above asked questions... ........

thanks a million in advance for giving answers for my above asked questions... .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

#4039

From: Manish Negi <tutor_best@...>
Date:: Sat Dec 19, 2009 5:39 pm
Subject:: Re: where we can use statistics.............

tutor_best

Offline

1] how can we implement statistics in Construction companies[Infra structure companies]

pert is used here which stand for project evaluation and review teachnique

2] How can we implement Statistics in Stores related works in any organization.

mode will be sufficient and queuing techniques........

3] how can we use statistics in Voluntary organization. ......... ...and social organization

to take the survey we use sampling techniques

--- On Sat, 19/12/09, kishore reddy <kishore_pingili@...> wrote:

From: kishore reddy <kishore_pingili@...>
Subject: [Statisticians_group] where we can use statistics.............
To: Statisticians_group@...
Date: Saturday, 19 December, 2009, 5:30 PM

Hello statisticians,

Where can we use statistics in our real life........ ....
i have some questions .............
1] how can we implement statistics in Construction companies[Infra structure companies]
2] How can we implement Statistics in Stores related works in any organization. ........
3] how can we use statistics in Voluntary organization. ......... ...and social organization
which statistical techniques we use in the above asked questions... ........
thanks a million in advance for giving answers for my above asked questions... .....

Rgds,
kshr

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.

The INTERNET now has a personality. YOURS! See your Yahoo! Homepage.