Teknik Jaringan Syaraf Tiruan Feedforward Untuk Prediksi Harga Saham Pada Pasar Modal Indonesia

To predict the condition of stock price, several technical analysis models have been used and expanded such as MACD, Fourier Transform, Accumulator Swing Index , Stochastic Oscillator etc. For input they are using the various prices such as Open, high, low , close , volume, BID, ASK price, and the output is a graphic that shows the decision whether to sell, buy or hold. Another method to determine the stock price by using Fundamental Analysis method. Fundamental method is an analysis that is based on the ratio or financial report from the existing company. Neural Network System Technology has been implemented in various applications especially in introduce the pattern. This power has attracted several people to use Neural Network for medical, Finance, Investment and marketing. Assuming that the prediction of the output system (next output prediction) is deterministic, than the suitable N.N model to predict it is Feed Forward. The prediction of the stock price is the complex interaction between unstable market and unknown random processes factor. The data from stock price can be determined by time series. If we have daily data from a certain period, for example : Xt(t = 1,2,...) than the stock price for the next period (t+h) can be predicted (the timing used can be in hourly, daily, weekly, monthly or yearly). To get the good prediction, the inputs from several aspects of the share prices have to be input in Neural Network after that the weighing principal can be adapted to minimize the wrong prediction in the first future steps. By using the final weighing, an action is done to done to minimize the total error in the second future steps. Due to that, the risk of Investor's decision to sell or buy the stock can be minimized. This paper will discuss on how to use and implement Time Series Neural Network to predict the stock market in Semen Gresik (SMGR) and Gudang Garam (GGRM