Predicting Trends of Stock Market Using SVM: A Big Data Analytics Approach

The huge quantity of data generated by several disparate data sources cannot be processed with traditional database tools and techniques. To analyze and extract events, patterns and useful information from such an enormous quantity of data requires high computational techniques. With embossing technologies like Hadoop and Apache Spark, it becomes feasible to treat this huge amount of data and extract valuable entities from it. In this paper, a stock prediction model is proposed using sentiments of tweets and news data. Classifier Support Vector Machine is implemented for trend prediction of 2 specific companies traded under the National Stock Exchange. A predictive model using Big data analytics has been utilized by considering millions of tweets and news data in Hadoop and Apache Spark framework. The output of the implemented model has been computed for the amalgamation of 3 types of input data. The paper compares the prediction output with existing researches implemented with SVM.