A Comparative Study of K-Nearest Neighbor, Naive Bayes and Random Forest techniques for Stock Market Prediction Model

Abstract

Nowadays, the stock market’s prediction is a topic that attracted researchers in different countries. Stock market prediction is a process that requires a comprehensive understanding of the data and analysis it accurately. Therefore, it needs intelligent methods to deal with this task to ensure that the prediction be as correct as possible, which will return profitable benefits to investors. The large number of companies traded in the stock market basket for various sectors makes it difficult for investors to predict the shares of a particular company or sector. This study aim discusses using the techniques of data mining for selecting the best model for forecasting financial stocks for companies. The study proposes to study three powerful forecasting techniques, namely Naïve Bayes (NB), Random Forest (RF), and K-Nearest Neighbor (KNN) for NYSE stock market data. The results of the experiments showed that Random Forest technique is best for companies’ stock prediction according to the error rate metrics, precision, recall, and F-measure.