|
Machine Learning Apply for Financial Credit Approval to Filter Selected Customer in Domain Specific Bank |
|---|---|
| รหัสดีโอไอ | |
| Creator | Uraiwan Inyaem |
| Title | Machine Learning Apply for Financial Credit Approval to Filter Selected Customer in Domain Specific Bank |
| Contributor | Sirina Chuaytem |
| Publisher | คณะวิทยาศาสตร์และเทคโนโลยี มหาวิทยาลัยเทคโนโลยีราชมงคลธัญบุรี |
| Publication Year | 2563 |
| Journal Title | วารสารวิทยาศาสตร์และเทคโนโลยี มทร.ธัญบุรี |
| Journal Vol. | 10 |
| Journal No. | 1 |
| Page no. | 36-46 |
| Keyword | Data Mining, Classification, Decision Tree, Financial Credit, Bank Domain |
| URL Website | http://www.sci.rmutt.ac.th/stj |
| Website title | Science and Technology RMUTT Journal |
| ISSN | 2229-1547 |
| Abstract | The objective of cooperative society is to support finance for members. A cooperative provides privilege of products and financial credit for members. The approval of the limit of financial credit is very complex condition. A recommending system can help cooperative officials to collect customer's profile and financial status to predict the limit of member's credit. The objective of the proposed research is to filter selected customer group for financial credit approval using data mining techniques. The dataset in the research with case study has 500 records used for preprocess transferring format into a categorization status value (CSV) file using WEKA program. The dataset is divided into 2 sets. First set is a training set used for creating model and the second set is a test set for evaluation model. For research methodology, the preprocessed data is a process using techniques of data cleaning to prepare data into suitable form before testing. The process of the research uses a data analysis technique which is Cross-Industry Standard Process for Data Mining (CRISP-DM). The experimental result is shown that accuracy value is 96.5517% with ADTree algorithm. The comparison between performance of algorithm are found that ADTree Algorithm and LMT Algorithm in test option, training set, shows maximum F-Measure 97.7% Precision 100% and Recall 95.5%. The proposed research is shown that the system can support cooperative officials to recommend the limit of financial credit for cooperative members in high quality. |
