This is a project for a contest hold by MMT, https://www.mmtvip.com/oversea, and Data Application Lab, http://datalaus.com/en/.
It is a Fintech model built to evaluate the credit scoring based on big data technology for international students in North America. The process for the application include: a) data preparation with web crawler, b) data Extract, Transform and Load (ETL) with MySQL, c) feature selection and d) model building with machine learning.
The model was built with decision tree classifier from scikit-learn package. Based on the data from Lending Club, the accuracy of the model is 25.8%. Compared with the random guess model, the model improves the accuracy by 11.5%.