/LSTM_Stock

Learing the process of LSTM, and use keras achieve stock prediction using LSTM。LSTM步骤解释,然后使用keras实现用LSTM预测股票走势

Primary LanguageJupyter NotebookMIT LicenseMIT

使用单层的LSTM网络实现对**股票市场的预测,使用股票市场中的其中一只作为训练模型,并对别的股票进行预测,一天对一天模型:趋势预测率达到90%

Using the single-layer LSTM network to predict the Chinese stock market, using one of the stock market as a training model and forecasting other stocks, one day to one day model:the trend prediction rate reaches 90%

结果:

Result:

一天对一天模型:

one day to one day model:

用训练集后30天预测:

predic using the last 30 days in train set:

test_set_ans

用最近30天预测:

predic using the recent 30 days:

related_day

用别的股票最近30天预测:

predic using the recent 30 days of another stock:

other

21天对3天的模型(非常不理想):

21 days to 3 days model(terrible):

predic using the recent 30 days of another stock:

21-3