財務模型智能分析與應用

The project aims to create a new type of asset management tool. According to the statistics company Statista's forecast, the US intelligent investment management assets will reach US$224.802 billion in 2017 and will reach US$509.55 billion by 2021. The annual compound growth rate will reach 29.3%. As of February 2017, the top five companies in the asset management category, Vanguard, Schwab Intelligent Portfolios, Betterment, Wealthfront, and Personal Capital, managed 470, 102, 73.6, 50.1, and $3.6 billion, respectively, and it can be seen that intelligent investment tools will be heavily utilized in the trend of future asset management. The plan integrates financial models, neural networks, and image recognition to create an AI system that can accurately analyze the future performance of "stocks". Our proposed methodology is different from the ones that used in the market in that we have employed image recognition to capture valuable information from outlook provided by analyst reports. The data mining/text mining of this system can be applied on a wide range of fields. Based on the empirical results, the number of investment decisions initially predicted by AI is still good, but it failed to significantly defeat the market. After further screening, its return on investment has increased significantly and defeated the return on the market during the backtest (24%).