DEVELOPING A CUSTOMER CHURN PREDICTION MODEL AND SENTIMENT ANALYSIS FOR PRODUCT EVALUATION AT PT X BY USING MACHINE LEARNING
PT X, one of education technology company, has a high percentage of churned customer for TryOut (TO) PTN product, at average of 73% based on last 5 TryOut. Churned customer is defined as customer who no longer uses a service or product (Hejazinia and Kazemi, 2014). As comparison, the tolerable churn...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/79361 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | PT X, one of education technology company, has a high percentage of churned customer for TryOut (TO) PTN product, at average of 73% based on last 5 TryOut. Churned customer is defined as customer who no longer uses a service or product (Hejazinia and Kazemi, 2014). As comparison, the tolerable churn rate for the education industry is 9.6%. It indicates a mismatch between expectation and actual service received by the customer. Thus, in this research, author implements prediction model for churned customer using data mining and machine learning techniques, using Cross Industry Standard Process for Data Mining (CRISP-DM) methodology. The selected model is built based on random forest algorithm. Sentiment analysis model is also built for this research, based on Bidirectional Encoder Representations from Transformers (BERT) algorithm and using text data from Twitter as input to know customer’s prespective about PT X. The result shows accuracy of 97% and recall of 100% for churned customer prediction model and accuracy of 88% for sentiment analysis model. |
|---|