Research of Quantative Trading Strategy Based on LSTM

Making strategies to maximize returns has always been the biggest problem almost every investor faces in process of investment. In this paper, we will establish a model analyze internal characteristics gold and bitcoin price data for further prediction, which is great help formulation trading strategy future. Section 2 hope trend bitcoin. At beginning based on ARIMA get relatively good prediction result. However, considering errors may occur when backward unit gets long model, try use LSTM algorithm better extract nonlinear two time series price. Through much result, thus choose as our final after comprehensive comparison. 3, predicted results make strategies. We first divide investors into reckless prudent types models under different assumptions respectively. Then by applying greedy establishing risk quantification approach global optimal solution end, testing confirming its excellent performance, summarize advantages shortcomings about comprehensively. Moreover, more conclusions are drawn possible update future.Price Prediction