Nepal Stock Exchange Prediction Using Support Vector Regression and Neural Networks

Abstract

Stock Exchange price prediction is the task of estimating future price of certain stock listed in stock exchange by extracting the trend with the help of confidence learned from historical training data. In this research work, the data set has been created by extracting raw data from Nepal Stock Exchange (NEPSE) website. Data preprocessing is performed in order compute an accurate result. The data belonging to promoter share and unwanted feature are eliminated from considered data. The resulting data are normalized for better performance, before applying the machine learning methods. Min-Max and Z-score normalization are used for this purpose. Overall stock data are further divided into ten different sector of investment for sectorwise analysis. Support Vector Regression (SVR) and Artificial Neural Network (ANN) are applied in order to predict stock price for a next day. In order to measure the performance of two learning models, mean square error (MSE), mean absolute error (MAE), root mean square error (RMSE) and Coefficient of Determination (R2) are used. The result shows that SVR with min max normalization is performing better than ANN in all sectors except on Development bank, Finance, and Mutual Fund.