Abstract
Today Stock market share typically considered to be a dynamic, non-linear, complicated, non-parametric, and chaotic in nature. The profitability of investing and trading in the stock market in large part depends at the predictability. If the trends of the share market is precisely forcasted , the shareholder may effectively be escorted and earn high return. Various statistical models are sufficient to analyse properly compiled, structured, and non-missing data. But, it is insufficient to analyse unstructured, noisy and improperly compiled data. So, statistical models for the prediction of stock market which is mostly based on data having certain patterns that might be noisy, unstructured and improperly compiled would be insufficient. Whereas, in many research, it has been established that, the most important capacity of the neural network is its proficiency to detect patterns and deformity as well as identifying multi-dimensional, non-linear co-relation in data.
Since the data of the Stock Markets are dynamic, non-linear, non-parametric, and chaotic in nature, so the prediction over such fluctuating data is not accurate with the help of conventional computers. Because, we cannot develop an algorithm, model, function or method to solve the problems based on such volatile data. But, the Neural Network helps us to get the solutions by formulating algorithm. Because, it follows different paradigm for computing based on the parallel architecture of Biological Brain. So, Artificila Neural Networks (ANNs) simply has the capacity to differenciate unknown and hidden structure in data which can be adequate for predicting the share market and stock exchange index. The Technical Analysis, Fundamental Analysis, Traditional time series forecasting are insufficient for stock market prediction. Hence, Standard Supervised Back-propagation Neural Network Learning Methodology where Boltzmann Machine, ADALIN, MADALIN will be used for the research.