Stock Market Prediction Using Hybrid Approach

Stock Market is becoming a new trend to make money. It is the fastest-growing system which is changing in every second. It is challenging and complex by nature which can make a drastic change in an investor's life. There are two possibilities either people will gain money, or he will be going to lose his entire savings. So for safe side stock market prediction is required, which is based on historical data. In this paper, we have proposed a hybrid approach for stock market prediction using opinion mining and clustering method. A domain-specific approach has been used for which some stock with maximum capitalization has been taken for experiment. Among all the available approaches our proposed model is different alike existing methods it not only considers general states of mind and sentiments, but it also forms clusters of them using clustering algorithms. As an output of the model, it generates two types of output, one from the analysis of sentiment while another one from clustering-based by taking popular parameters of stock exchange into consideration. The final prediction is based on an examination of both the results. Also, for empirical analysis, we have considered stocks with maximum capitalization from 6 growing sectors of India like banking, oil, IT, pharma, automobile, and FMCG. As a result, we have observed that predicted values from the proposed approach show maximum similarity with the actual values of the stock. The hybrid model returns efficient results in terms of accuracy in comparison with other individual methods of sentiment analysis and clustering.