Machine Learning Based Stock Market Trend Prediction and Analysis

Abstract

Stock price prediction can help investors to create initial pre-scenarios. The topic of this research is to predict the stock market scenario through machine learning methods. By successfully predicting the stock market situation, the movement of different stocks, etc., the possibility of buying the wrong stocks can be greatly reduced, making it possible to make huge profits by buying and selling stocks. The purpose of the research is threefold. This paper uses a market capitalization weighted index consisting of the most important 40 stocks out of the top 100 stocks with the largest market capitalization on the Paris Stock Exchange as a dataset to analyze the movement of stocks and the possibility of buying them. The stocks are analyzed and subsequently forecasted through data visualization techniques such as candlestick charts and moving average charts. Long Short-Term Memory (LSTM) was used as a benchmark. This study has three main objectives: first, to study the temporal evolution of the closing prices of these stocks; second, to make stock forecasts using candlestick charts, which is the traditional method of stock analysis; and third, to analyze the closing prices of the stocks and the moving average charts, which are studied and evaluated.  Stock prediction can reduce risk, guide investment decisions, realize risk management, make market references, assess the value of stocks and enhance strategic vision.