Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

Final Project MET AD654

Deliverables:

· Group Presentation (15 minute presentation of process, results, and steps)

· Two Page Summary Memo

· Answer, comments, and code related to steps below.

The HBS Case “Improving Lead Generation at Eureka Forbes Using Machine Learning Algorithms”  is a strong parallel to the types of analytics problems Marketing Analysts face in their daily work.  You will simulate this type of work by putting yourself in Kashif’s shoes.  

1) Read the HBS Case “Improving Lead Generation at Eureka Forbes Using Machine Learning Algorithms”

2) Discuss the case as a team and perform the next three steps together before beginning your analysis.

3) Perform a 5 Forces analysis using the information available in the case. Summarize the competitive aspects of the market.

4) Perform “Ps” analysis on Eureka Forbes.  Use any Ps you think are appropriate, including any of the new Ps that you think are relevant.

5) Do you think this would be a high HHI industry?

6) Explain the problem that Kashif is working as an Analytics question.  What is he trying to understand, and what are the objectives of his analysis?

7) What are key components of CAC for Eureka Forbes?

8) Use a chi-square test to analyze the following claims:

a. Converted customers using mobile, desktop, and tablet are equally distributed

b. Repeat visitors are as likely to convert as new customers

9) Use a two-sample t-test to analyze the following claim:

a. Customers who convert spend more time on the website

10) Build a confusion matrix for a logistic regression model on the given data.

11) Build a logistic regression model on the data.

12) Interpret the results.

13) Construct a simple decision tree (classification tree).  Explain what the disadvantages are of using decision tree classifiers.

14) What are the AUCs in your decision tree? Is there any overfitting?

15) Explain how this data set is imbalanced.   What considerations can be made to improve prediction in these cases?

16) Examining your Logistic Regression Model, and Simple Decision tree, describe the disadvantages when the classes are not balanced?

17) The data has a large number of variables and is highly imbalanced.  Do you think techniques such as logistic regression can be applied when the number of variables is large?  

18) What variable reduction techniques can be used to make the model more efficient?

19) Use the sample data provided and appropriate sampling data to develop a random forest model.  Describe in detail how are you approaching the model and its objectives.  

20) Consider a SMOTE sampling model and see if you can improve your predictions.

21) Tune your model based on parameters, accuracy, and sampling approach.

22) Comment on the model development and accuracy of the model.

23) Going back to point #4, what suggestions do you have for changing the A-I-D-A approach of Eureka Forbes to improve profitability?   

24) Summarize your findings as a 2 page memo addressed to Shashank Sinha, the Chief Transformation Officer of Eureka Forbes. Your memo should be the most polished part of your deliverable and explain your findings.   A successful memo will describe your approach to the analysis, findings, and key recommendations in a concise manner.