A Financial Forecast Model based on LSTM Algorithm and Goal Planning

Trading strategies play a very important role in financial asset trading, and obtaining ideal trading returns in complex and volatile financial markets is a research hotspot. In this paper, a hybrid transaction model based on LSTM algorithm (Long Short-Term Memory Network) and Goal programming model is designed, and the optimality of the model is proved by the similarity ranking preference technique of ideal solution (TOPSIS). In this paper, the LSTM model can predict the trend of the price time series, and then use the goal planning model within a certain risk range to maximize the trading income in a certain period. Through the verification analysis, the final rate of change of the return rate in the experiment of slightly changing the commission rate is less than 2%, which proves that the model introduced in this paper is robust to a certain extent.