A New Model for Stock Market Predication Using a Three-Layer Long Short-Term Memory

Abstract

The stock market predication is full of uncertainty and is affected by many factors. such as historical data, sentimental analysis, and financial analysis. Studies have also shown that predicting direction taking only one factor or merging two factors only from three different factors (Trend analysis - Financial ratios - Sentiment Analysis). A new proposed model will be used, which combines three features of stock evaluation for the first time: trend analysis based on historical data, analysis of financial ratios, and online news analysis as an input to the deep learning algorithm using long short-term memory to get more accurate results. It can be shown from the results that this model can be used to predict the stock performance with a higher accuracy compared with other models.