Performance of Prophet in Stock-Price Forecasting: Comparison with ARIMA and MLP Networks

Prophet, an open-source forecasting tool developed by Facebook, is increasingly being used to predict stock prices. This study compares Prophet with two other popular forecasting tools: ARIMA and Multi-layer Perceptron (MLP) networks. Prophet is particularly precise at identifying key turning points and adjusting for holidays and outliers, making it suitable for capturing long-term trends. On the other hand, ARIMA is better suited for short-term predictions, especially in cases with stable, linear trends. Meanwhile, MLP networks excel in managing complex datasets and capturing strong nonlinear patterns. In our analysis, we used data from Turkey's BIST 30 index and a selection of 100 corporate stocks from Yahoo Finance. We evaluated the accuracy of each method by comparing the predicted stock values against actual outcomes using correlation coefficients. The goal of this study is to determine how well Prophet performs in forecasting stock prices compared to these two commonly used methods.