├── CustomerChurn_ModelBuilding_Validation.ipynb
├── CustomerChurn_dataProcessing.ipynb
└── README.md


/README.md:
--------------------------------------------------------------------------------
 1 | ## Customer Churn-Case-Study
 2 | 
 3 | 
 4 | #### Entire case-study cannot be uploaded as it is copyrighted
 5 | 
 6 | ABC store has an online portal where users can purchase certain products such as A and B and the latter is the important driver of revenue generation. Major source of revenue comes from renewal via upgrade to the latest software but this year no upgrade is planned for workstation.
 7 | The business challenge is to target new workstation customers, up-sell to existing customers who have not upgraded to the available latest version and cross sell to the customers who don’t have workstation.To increase the sale, ABC would target potential customer who are likely to buy by offering them coupons.
 8 | Now looking at the problem from analytics angle, ABC has rich source of information about the customers including their purchase history and email ids. As the customer base is already high, it is not advisable to send the coupons straightaway to all the customers due to budget constraints.
 9 | Now using analytics and  machine learning techniques, analytics team can build a propensity model to identify the customer who are likely to buy. Lift/Gain chart could help in identifying the targeted audience
10 | 
11 | **Major challenges**
12 | * Highly imbalanced Dataset
13 | * Large dataset
14 | 
15 | **Data Balancing techniques used**
16 | * Random Sampling
17 | * SMOTE:Synthetic Minority Over-sampling Technique
18 | * SMOTEEN
19 | * Cluster Based Over sampling
20 | * Bagging
21 | * Boosting
22 | 
23 | **Feature Selection Techniques**
24 | * Extra Tree Classifier
25 | * Wrapper Methods
26 |   * Recursive Feature Elimination
27 |   * Forward Backward Selection
28 | * Variance Inflation Factor
29 | * Embedded Techniques-LASSO & RIDGE
30 | * Boosting Techniques-Gradient Boosting Tree
31 | 
32 | **Algorithms Implementation**
33 | * Logistic Regression
34 | * Regularized Logistic Regression
35 | * Random Forest
36 | * Support Vector Machine
37 | * Adaboost
38 | * XGBoost
39 | * Lasso & Ridge
40 | 
41 | 
42 | 
43 | 
44 | 


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