Enhancing stock prediction clustering using K-means with genetic algorithm

Currently, the market has been facing many rapid changes and challenges, particularly with social media outlets affecting the market liquidity, but also helping most researchers in generating predictions data from commercial applications to overcome the unpredictability of the stock market. Twitter and Facebook act as two of the most important sources to extract data from, as well as good examples for how social media data reveals great impact on the public and their future behavior. This research tries enhance the previous “An Intelligent Framework Using Hybrid Social Media and Market Data, for Stock Prediction Analysis” [1]. Through investigation, it was found that previous results were not promising and did not achieve the investor's satisfaction. Therefore, cluster algorithms were developed by combining genetic algorithm and k-means. The main objectives of this research are to optimize the clustering of stock market prediction and to examine the impact of applying genetic algorithm optimization with k-means clustering algorithm. The objectives were approached by using Chi-Square similarity measures for accuracy and the sum of square distances (SSD) of the selected clustering algorithm. The evaluation shows that using genetic algorithm and k-means clustering algorithm with Chi-square similarity measure achieved the highest accuracy with the least sum of square distances.