Short-Term Stock Price Prediction Models Based on Economic Background

In recent years, many different types of prediction models have occurred, and many of them are utilized in the stock market, to predict the stock price for a relatively short period of time, to speculate. Nonetheless, they have both advantages and disadvantages, some models are only reliable to a certain extent, or some specific conditions. In this case, predicting the fluctuation of the stock prices based on a single model possibly lead to an inaccurate value, and this paper aims to analyze the weakness and benefits of several models in these three categories, including pure Statistics, time series, and deep learning. And this article purposes to collect historical data, and utilize the data to analyze different models from the three categories, study the performances of different models in the various economic background, refer to the realistic macro data in different periods. Deriving a new method, giving the consequences from the models' new meanings according to the economic background, is the innovation of this study.