Research On The Application Of Improved Convolutional Neural Network In Financial Forecasting

With the development of the world economy, the world finance is in a stage of rapid development and the financial activities are increasing. The uncertainty of the change trend of financial activities is also increasing. How to learn and master the rules of financial activities and predict the future change trend of financial activities become the focus and the main research content of academic and financial areas. Financial prediction can effectively provide the basis for making financial plans and decisions, and then maintain the healthy development of the financial market and maximize the profit of financial organizations.Convolutional neural network is a typical deep learning model. It comes from the structure of the human brain’s visual system. It is a kind of multilayer neural network which consists of convolution layers and sampling layers. It can learn effective features from a large number of input data, and the higher-order representation it learned contains the input data structure information. It is a good way to extract features from the data. Currently convolutional neural network has been widely used in speech recognition, image recognition, natural language processing and other fields. Therefore convolutional neural network is introduced for predicting financial time-series data in this paper.In this paper, the domestic and foreign research methods on financial time series have been reviewed and summarized. The artificial neural networks and deep learning methods are introduced briefly. Besides,convolutional neural network and support vector machine algorithm are introduced especially. The main works in this paper are as follows:According to the characteristics of financial time series data, the convolutional neural network is improved, and a convolutional neural network forecasting model is established. Then the model is applied to stock index prediction. The influence of model parameters on stock index prediction is studied;To improve the convolutional neural network stock index prediction model, the advantages of the effective feature extraction with the convolutional neural network and the good classification and prediction ability of the support vector machine are combined. A hybrid prediction model based on convolutional neural network and support vector machine is proposed to predict the stock index, which can improve the prediction accuracy;The exchange rate forecasting model based on the convolutional neural network is established, and the influence of the model parameters on the exchange rate prediction is studied. The model of the hybrid model of the convolutional neural network and the support vector machine is used to predict the exchange rate. The feasibility and effectiveness of the two proposed models is verified by comparing the simulation results.