Research on Predicting Credit Card Customers’ service using Logistic Regression and Bp Neural Network
Download as PDF
As the number of credit card customer increases, the urgency for an efficient model to predict customer services became larger than ever. On the other hand, machine learning is also growing as a new branch of artificial intelligence. Its ability to process large amounts of data sets and do tasks with human monitoring well suits the market of predicting customer services. However, machine learning is sorted into many different kinds of models with different advantages and drawbacks. The Logistic regression model is used to analyze and explain the relationship between a nominal scale response variable and more than one explanatory variable. In addition, Logistic regression model is the best for credit card companies because of its clear and easily interpreted outputs. However previous research has shown that neural networks seem to have a higher precision and accuracy. In order to test for the truth, a comparison of Bp neural network and Logistic regression model is being made based on the data set provided by Kaggle. The data set included 23 kinds of information about 10128 customers and outliers had been previously eliminated.
Machine Learning, Data science, Logistic Regression, Bp Neural Network, credit card service