Time series forecasting of stock market using ARIMA, LSTM and FB prophet

Considering turbulent character, predicting the future stocks of a company is a difficult endeavor. The goal of this study is to analyze the performance of three widely used forecasting methods: ARIMA, LSTM, and FBProphet. ARIMA is a time series data statistical model that captures linear relationships and stationarity. Recurrent neural networks, such as LSTM, are able to recognize nonlinear patterns and long-term dependencies. FB Prophet is a Facebook-developed time series forecasting library that use an additive regression model to account for trend, seasonality, and holiday impacts. The results show that each strategy has advantages and disadvantages in projecting stock market values. When the underlying data is steady and linear, ARIMA works well. In contrast, LSTM excels in capturing nonlinear and complicated relationships. FB Prophet performs admirably when dealing with trend and seasonality patterns.This study examines the performance of ARIMA, LSTM, and FB Prophet in stock market forecasting, allowing practitioners to choose the best approach depending on the peculiarities of their data and forecasting objectives. Further study might look at ensemble methods or hybrid approaches that combine the capabilities of these techniques to increase stock market forecast accuracy.