Forecasting Methods of Big Data Based on Machine Learning and Cloud Computing

The history of credit cards could date back to 1950s. Credit card gradually become one of the main forms of payment for people. In this case, the default risk has become one of the biggest problems in credit risk and how to identify this risk is more and more important. This report aims to solve the problem about identification of default risk in bank before the credit extension or other service. The main solutions for this problem at present are setting black and white list, fixed threshold filtering. However, these two solutions have a lot of drawbacks: 1. They cannot be used to deal with a large number of loan transactions. 2. They take a lot time to complete estimation, which is not efficient. 3. The accuracy of these methods is low. Based on these drawbacks, I put forward methods by using machine learning to predict whether the borrowers will default.