Skip to content

Latest commit

 

History

History
37 lines (25 loc) · 613 Bytes

README.md

File metadata and controls

37 lines (25 loc) · 613 Bytes

Customer-Churn-Prediction

The purpose of this research is to develop and design an effective and efficient model for customer churn prediction in telecommunication industry

Data Preparation Steps

  1. Data Conversion

  2. Sparseness Elimination

  3. Outliers Detection

  4. Features selection

  5. Data Encoding

  6. Data Standardization

Implemented Models

  1. Gradient Boost
  2. KNN
  3. xgBoost
  4. Decision Tree
  5. AdaBoost
  6. Random Forest
  7. Logistic Regression
  8. Gaussian NB
  9. SVM

Evaluation

  1. Trains and test scores
  2. Accuracy
  3. Classification Report
  4. Confusion Matrix
  5. Area under curve