COMPARISON OF STOCK PRICE PREDICTION ACCURACY WITH VARIATION NUMBER OF HIDDEN LAYER CELL IN BACKPROPAGATION ALGORITHM

Abstract

Nowadays there are various ways to invest money. The capital market becomes a tantalizing choice for potential investors. The profit from this investment is obtained from the difference between the selling and buying price of the shares. Created program is expected to provide an overview comparison of prediction accuracy from different hidden layer cells amount. Using historical data from yahoo finance to get stock prices in last 12 months. The data is processed using backpropagation algorithm and the process is done by the number of different hidden layer. Eventually the program will give the stock price predictions from four different number of hidden layer. Of all the proceeds will be matched to determine which one has the most precise accuracy.