Abstract - Neural networks as an intelligent pattern recognition systems are used for prediction of Stock prices.However,there is no exact method to form an architecture to predict exact price of stocks.In this paper, a feed forward Multilayer Perceptron (MLP) and Support Vector Regression are used to predict a company's stock value based on its stock price value of history.
Introduction - Stock trading is one of the most important activities in finance.This paper simply tells the method to predict the future price of stocks to make educated decisions on investments.In recent years, the immense popularity and accuracy of machine learning in various industries have enlightened many traders to apply machine learning techniques , and many of them have shown promising results. In this paper, we have focused on short-term price prediction on the price of stocks.It is really difficult for the human beings to process such a large amount of data and make prediction on it.It is possible in recent years only by the fast development of computers.
Research studies have shown that Neural Networks have great capability in pattern recognition and machine learning problems such as classification and regression.Neural Networks are self-adjusting methods based on training data, so they can solve the problem with a little knowledge about its model. Besides, neural networks can find the relationship between the input and output of the system even if this relationship might be is very complex because they are generally function approximators.
Most importantly,Neural networks have generalisation ability so that they can predict the results even on the data on which they are not even trained.The basic idea for using neural networks for predicting problems was first used by Hu in 1964 for weather forecasting [1]. The absence of learning method for multi layer networks made it very difficult to apply these networks to complex prediction problems. But in 1980s the back propagation algorithm was introduced for training a MLP neural network. In recent time many researchers have used neural networks to predict the stock market changes. Kimmoto and his colleagues used this method in which they used neural networks to predict the index of Tokyo stock market [2].Many methods can be used to predict stock prices and can be compared[3].Here,we have suggested a predictive model based on MLP and SVR neural network for predicting stock market changes.