Stock market is a high risk investment influenced by many factors. Stock market prices prediction is not an easy task. With the aid of classification, a data mining technique predicting stock prices considering some factors of influence had been done. This paper put a light on the performance of ID3 and Naïve Bayes algorithms on a given Stock market data. ID3 and Naïve Bayes were classification algorithms which classifies the given data to be classified (test data) basing on the historical data (training data) provided. The historical and test datasets contain attributes which are the factors influencing the stock prices. ID3 algorithm is a Decision Tree technique which constructs a decision tree using the historical data. After decision tree construction, prediction is done for the test dataset values and forecast accuracy is calculated using original value dataset values. Bayesian networks are also used for prediction. The Naïve Bayes algorithm is a Bayesian Network technique used for the Bayesian Network construction using the historical data. The constructed Bayesian Network aids in prediction of the test dataset stock prices and forecast accuracy is calculated using original value dataset values. For computing forecast accuracy root mean square deviation is used. Along with forecast accuracy, under and upper forecasting of the algorithms are also presented. These two algorithms namely ID3 and Naïve Bayes are evaluated on various stock market datasets and the comparison of their performance is provided.

Stock Market Decision Trees ID3 algorithm Bayesian networks Naïve Bayes algorithm comparative analysis.