Real Time Stock Movement Estimation Using Data Mining and Soft Computing Techniques

Abstract

 Predicting the stock prices with the use of machine learning and neural networks isn't a new approach, but having it as a huge data that is volatile and unpredictable even to some of the best algorithms. As the stock data is purely historic using only this to predict the values is not efficient. So, applying some of the basic technical analysis methods to the data before passing through the neural net would make the task of finding the market patterns easier. This would also remove the human error in finding the relations in the technical analysis and speed up the process. This algorithm would be scalable as any new approach can be implementedto the data before sending it to the network.