Stock Market Prediction Accuracy Analysis Using Kappa Measure

Abstract

The nature of stock market is highly stochastic which can only be predicted. There are various companies and news channels which uses different data analysis tool to forecast the behavior of the stocks on day to day basis. They also provide tips and recommendations to buy and sell certain stocks which will lead to more profit. As there are many news channels, websites and organizations which are doing this, it is very difficult for the buyer/seller, to judge whom to believe and whom to ignore. In this paper, we have applied kappa measure to quantify the accuracy of stock market prediction by various media houses. The stock with the highest kappa measure can be considered to be the best stock to buy. Moreover, Kappa measure also finds the risk involved in the purchase/sale of each shares. Thus instead of believing on a particular channel, newspaper or website for the stocks that should be purchased/sold, its combinations are used which improves the confidence in stock market recommendation.