Real-Time Internactive Investment Strategy Visualization System Using Deep Learning

• Main Authors: TSENG,SHAO-HUNG, 曾少宏
• Other Authors: HUANG,HSIAO-YUN
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/d62zw8

碩士 === 輔仁大學 === 統計資訊學系應用統計碩士班 === 107 === Now is an era of information explosion, and data is constantly being produced in large quantities. In this case, how to find useful information in such a large amount of information is an extremely urgent and important issue. The financial market is an area of Real-Time change. If only the analysis method is used for these data, some information or some analysis results may be ignored. Generally, the public can't easily understand the content, and information visualization is a kind of public. A clear-cut approach. In the past, the visualization of the financial system research, the chart is less interactive; In terms of data, most of the historical day data is used and the regular investment method is used to conduct stock backtesting through historical day data to judge the timing of buy and sell and the rate of return calculation; In terms of system, Python Django has not been or is rarely seen in related research.   Therefore, this research uses the minute data of stock market trading, using the integrated autoregressive moving average model, deep neural network and long-term and short-term memory to compare stock price forecasting and strategy evaluation. Finally, bulid a visualization system have interact and reat-time features use Python Djange. Provide a reference for public. After research result, the overall prediction accuracy of the LSTM model is better than ARIMA and DNN models. . The main functions of the system include descriptive statistics, stock price volatility, channel line and OHLC charts of four technical indicators in stock overview. In real-time stock price analysis, there are functions of real-time quotation, model prediction and research results.