Innovative Financial Prediction Model Based on Lasso and Keras Neural Network

摘要

Data analysis has been widely used for mining valuable information. This paper first studies the influence factors for the data of city revenue, and then mines the hidden relationship utilizing the Lasso regression analysis of selected key features. Finally, we construct our innovative financial prediction model based on Lasso and Keras neural network (PMBLanK). By comparing the performance of our PMBLanK model with other widely used prediction models, we can see that our PMBLanK model has higher performance over existing models with lower MSE, which can be used to predict fiscal revenue of coming years effectively.