PUBLIC COMPANY BANKRUPTCY PREDICTION MODEL DEVELOPMENT WITH DATA MINING TECHNIQUE
Public company bankruptcy cases have big impact toward many stakeholders including the company, investor, creditor, consumer and government. Bankruptcy prediction become an important tool in early detection whether a company will become bankrupt or not. This thesis test whether usage of data mining...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/43689 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Public company bankruptcy cases have big impact toward many stakeholders including the company, investor, creditor, consumer and government. Bankruptcy prediction become an important tool in early detection whether a company will become bankrupt or not. This thesis test whether usage of data mining technique based on CRISP-DM will generate prediction models with prediction performance better than prediction method in economics like Altman Z-Score. The experiment consists of testing combination of sampling method and modelling method. The experiments are done 3 times which first is using Altman variable as feature, feature adapted from Polish dataset and features selected using feature selection from the feature adapted from Polish dataset. The experiment result show that the performance of bankruptcy prediction models that was generated are better than Altman Z-Score performance. Usage of multiple feature beside Altman feature generate model that are decent but not better than using Altman Feature. |
|---|