Novel Approach for Stock Prediction Using Technical Analysis and Sentiment Analysis

Stock prediction is not new; people have tried to predict stock price or good quality stocks for ages. Machine learning has opened a new direction for the problem of stock prediction. The critical factor deciding the success or failure of the machine learning model depends on the quality of features computed. Moreover, the concept of ensemble learning can significantly enhance the quality of prediction. The proposed model gathers data price time series data and news articles from open access domain and do necessary pre-processing. The model also calculates technical and sentiment features and uses them for training the ensemble model. Performance is compared to definite matrices and with other similar research.