kurofufu Tracker

a regression model question

kurofufu — Fri, 01 Feb 2013 01:39:59 GMT

The monthly sales amount of a retail store(S) is positively correlated with the number of visiting customers(C). The regression model of S on C has a positive regression coefficient for C. But after I add some additional variables, the reg coefficient for C continues to be significant but becomes negative. Should I include C in the model? If so, the interpretation of C effect will not make sense at all. Any ideas?

Statistical Testing for A/B test, MVT

kurofufu — Fri, 04 Jan 2013 06:51:04 GMT

A/B testing and Multivariate Testing(MVT) are popular testing methods used in web analytics. Can anyone tell me which statistical procedures I should use for these testings?

How to get ROC curve in Model Comparison

kurofufu — Wed, 02 Jan 2013 19:20:17 GMT

I run model comparison node and open the results window, but only see Score Distribution, Score Rankings, Fit Statistics and Output windows. I would like to know how to open ROC curve winodw.

Re: Regression question

kurofufu — Mon, 12 Nov 2012 19:00:55 GMT

I just checked. Three regression have the same number of Ns and no missing values in X1 and X2.

Re: Regression question

kurofufu — Fri, 09 Nov 2012 21:54:36 GMT

The fitted model using X1 and X2 is

Y = 0.5 + 1.2*X1 + 1.1*X2 + e

This equation tells me that if I increase one unit in X2 while X1 remains constant, Y will increase 1.1. Is this a wrong conclusion given that when Y is regressed on X1 only, X1 is not significant?

R replaces SAS/STAT?

kurofufu — Fri, 09 Nov 2012 21:48:09 GMT

Will R replace SAS/STAT someday? Now big internet companies such as Google, Facebook etc only look for people with R skills.

Re: Regression question

kurofufu — Fri, 09 Nov 2012 21:44:53 GMT

The adjusted R-square is the same - R-square = 0.7598, Adj R-square = 0.7581

The interaction term X1*X2 is NOT significant.

Regression question

kurofufu — Fri, 09 Nov 2012 21:14:14 GMT

If Y is regressed on X1 only, R-square is 0.72 and the X1's coefficient is significant.

If Y is regressed on X2 only, R-square is 0.002 and the X2's cofficient is NOT significant.

If Y is regressed on X1 and X2, R-square is 0.76 and the coefficients of both X1 and X2 are significant.

Questions:

1. Why X2 becomes significant when it is combined with X1? Is there a more intuitive explanation for this?

2. Should I use X2 in the final model?

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 18:33:02 GMT

thanks for the great answer, pagemiller.

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 18:01:58 GMT

Ok, let's focus on predictive modeling for this question.

When we create a regression model for prediction, don't all coefficients included in the model need to be significant?

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 17:45:32 GMT

GLM also produces the same result as REG.

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 17:12:32 GMT

Sir, we are creating a regression equation for prediction, not for comparison of two groups.

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 17:03:51 GMT

Don't you need to report significance information when presenting a regression equation?

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 16:48:10 GMT

I understand c0 not equal to b0 and two equations are equivalent agebraically. But since c0 is not significant, how can we adopt the second equation?

Re: Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 16:32:22 GMT

Use regression equations to explain my question

First case: Y= b0 + b1 * X1 + ....

2nd case: Y = c0 + c1 * X1 + ...

b0, b1 are significant

c0 is not significant, c1 is signficant

So which one I should use? Which one is correct?

Indicator Variable in PROC REG

kurofufu — Fri, 26 Oct 2012 06:53:20 GMT

I create a indicator variable X1 to set 0 for group A and 1 for group B and then run proc reg on X1 along with other continuous X variables and both intercept and X1's coefficient are significant. But if I code it as 1 for group A and 0 for group B, then intercept becomes not significant. So what happens here? How to explain this?

Re: How to test a marketing effect?

kurofufu — Fri, 19 Oct 2012 18:41:08 GMT

suppose the data look like below and month 1, 2,3 are before the marketing campaign and month 4, 5, 6 doing marketing campaign.

region, store#, month, sales

A, 001, 1, 100

A, 002, 1, 102

....

B, 101, 1, 67

B, 102, 1, 105

....

A, 001, 4, 99

A, 002, 4,110

...

B, 101, 4, 105

B, 102, 4, 121

....

How to test a marketing effect?

kurofufu — Fri, 19 Oct 2012 18:10:31 GMT

Suppose a company has retail stores in two regions, A & B and ran a marketing campaign in region B for 3 months.

The average monthly store sales of last three months BEFORE the campaign of A is 5% higher than those of B.

The average monthly store sales of last three months AFTER the campaign of A is 8% higher than those of B.

I would like to know how to test 8% is significantly higher than 5%.

Re: surveyselect question

kurofufu — Sat, 22 Sep 2012 04:47:07 GMT

Any comments on this paper for my question?

http://support.sas.com/resources/papers/proceedings10/213-2010.pdf

Second, can we use propensity matching here?

Suppose we have 2000 members available for creating a control and test groups.

1. Randomly select one members from them

2. Use propensity score matching to find a matching member

3 Repeat step 1 and 2 until exhaust all members

ps. SAS paper for propensity score matching: http://www2.sas.com/proceedings/sugi26/p214-26.pdf

Re: surveyselect question

kurofufu — Tue, 18 Sep 2012 03:41:13 GMT

Thanks! This saves me a lot of time.