Application of Improved Convolution Neural Network in Financial Forecasting

Abstract

The world economy is in a stage of rapid development, and financial development is also continuing. Financial activities are increasing, and the uncertainty of its changing trend is also increasing. An effective financial forecast can provide a basis for financial planning and decision-making while maintaining the healthy development of financial markets. Convolution neural network is a multilayer neural network structure that simulates the operation mechanism of biological vision system. It is a neural network composed of multiple layers of convolution layers and down sampling layers. It can obtain useful feature descriptions from raw data and is an effective method for extracting features from data. Therefore, this paper introduces the convolution neural network structure to predict the financial time series data, establishes a convolution neural network model, and studies the influence of model parameters on the stock prediction results. Through simulation and comparison, the feasibility and effectiveness of the prediction model given in this paper are verified.