urbanclimatefr / telecom-customer-churn-prediction

Supervised learning algorithm was used to build churn prediction model to help solve a telecoms company's customer churn problem.

Telecom Customer Churn Prediction

• The source files consist of 2 parts.

• The first part is 2022_02_20_churn_summative_part_1.ipynb

• The input file required for part I is cell2celltrain_Small_6k.csv

• The second part is 2022_02_26_churn_summative_part_2.ipynb

• The input file required for part II are:

• 1. df_imputed.csv
• 2. features_selected_new.txt

• The formal written report is report.pdf.

• The requirement of this project is in Assessment Brief.pdf.

• Supervised learning algorithm was used to build churn prediction model to help solve a telecoms company's customer churn problem.

• Decision tree classifiers and optimisation techniques were used for feature selection.

• The genetic algorithm was applied to a telecoms customer dataset consisting of 6380 rows and 57 features.

• The Python programming language, Jupyter notebook and scikit-learn python package were used.