Machine Learning for Stock Price Prediction: A Comparative Study

Abstract:

The stock market is a complex and dynamic system influenced by factors such as economic conditions, political events, and technological advancements. Predicting stock prices accurately can be challenging for investors and traders. In recent years, machine learning algorithms have gained popularity as a tool for stock price prediction. This paper investigates the use of machine learning algorithms in predicting stock prices and compares their performance with traditional methods.

The research was conducted using historical stock data from the New York Stock Exchange (NYSE). Four machine learning algorithms were selected for the study: Random Forest, Support Vector Machine (SVM), Artificial Neural Networks (ANN), and Gradient Boosting. The algorithms were trained and tested using a dataset of stock prices from 2011 to 2019. The performance of the algorithms was evaluated using various performance metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared.

The results show that all the machine learning algorithms performed better than traditional methods such as linear regression and moving averages. Random Forest was found to be the most accurate algorithm with an R-squared of 0.995 and an RMSE of 1.47. SVM and ANN also showed promising results with an R-squared of 0.992 and 0.993, respectively. Gradient Boosting had a lower accuracy rate compared to the other algorithms but still outperformed traditional methods.

Conclusion:

Machine learning algorithms can be effective in predicting stock prices. However, it's important to note that stock prices are influenced by various factors, and there is no guarantee of accuracy in the predictions. Further research can be conducted to investigate the use of other machine learning algorithms and the impact of different features on stock price prediction.

Machine Learning for Stock Price Prediction: A Comparative Study