Welcome to the Simplilearn Community

Want to join the rest of our members? Sign up right away!

Sign Up

Attrition Analysis ( Datascience with SAS)

_26245

Member
Hi Priyanka,

I am working on Employee Attrition Analysis project. Can I use proc Logistic to check Max and Min values for the probability of churn. Also, what analysis can I determine using Proc Univariate . I need help with the last two points.

Here is my code:

title 'Stepwise Regression on the probability of churn';

proc logistic data=work.attrition outest=betas covout ;

MODEL Retain_Indicator (event='1') = Sex_Indicator Relocation_Indicator Marital_Status

/ selection=stepwise

slentry=0.4

slstay=0.45

details

lackfit;

Output out = outdata p=PREDICTED lower=LCL upper= UCL;


RUN;


PROC UNIVARIATE DATA = WORK.Attrition;

VAR Retain_Indicator;

Histogram Retain_Indicator/Normal;

RUN;
 

Priyanka_Mehta

Well-Known Member
Simplilearn Support
Hi Priyanka,

I am working on Employee Attrition Analysis project. Can I use proc Logistic to check Max and Min values for the probability of churn. Also, what analysis can I determine using Proc Univariate . I need help with the last two points.

Here is my code:

title 'Stepwise Regression on the probability of churn';

proc logistic data=work.attrition outest=betas covout ;

MODEL Retain_Indicator (event='1') = Sex_Indicator Relocation_Indicator Marital_Status

/ selection=stepwise

slentry=0.4

slstay=0.45

details

lackfit;

Output out = outdata p=PREDICTED lower=LCL upper= UCL;


RUN;


PROC UNIVARIATE DATA = WORK.Attrition;

VAR Retain_Indicator;

Histogram Retain_Indicator/Normal;

RUN;
Hi, Good effort in resolving the problem statement..

I would liek to inform you that, definitely you can use the Logistic regression to find out the min and max probability of churn. I am sure you are referring to a mentoring session where we have discussed the same. So just to notify you again that always remember, there are many ways to achieve a particular task and being a data scientist, you should always explore different ways to do the analysis.

Now regarding, Univariate, it is used to request a variety of statistics for summarizing the data distribution of each analysis variable: It is used to do descriptive statistics for multiple variables. I hope this will help you.
 

_26245

Member
Hi Priyanka,

I am working on Attrition Analysis project. the p value of all the variables are greater than 0.05 hence not Significant . in that case are we suppose to drop all variables from the model.

Thanks for your time and help.
 

Priyanka_Mehta

Well-Known Member
Simplilearn Support
Hi Priyanka,

I am working on Attrition Analysis project. the p value of all the variables are greater than 0.05 hence not Significant . in that case are we suppose to drop all variables from the model.

Thanks for your time and help.
Hi,

Apologies for the delay here.
This to inform you that, yes the variables here are not significant as we have a small data. So we will drop the variables. However, the purpose of this project is to give you an exposure to the concepts studied in the course.

I would suggest you that, practive with more datasets and implement your learning for a detailed exposure. You can refer the repository of datasets i.e. Kaggle.com. You will get a great exposure and a good hands-on.
 
Top