Abstract

Based on the market-efficient hypothesis, stock prices follow a random patrolling process. In such a market, stock returns cannot be predicted based on past price changes. However, providing evidence of the existence of exceptions in the stock market by researchers has added some uncertainty to the market-efficient hypothesis. Less reaction to the information is one of these exceptions on which the momentum strategy is based claiming the continuation of the stock price trend. Accordingly, in this research, it has been tried to use the capital market indices to represent the market trend and in the form of a neural network to analyze the time series of stock prices and the overall, financial and industry indices. This study is a descriptive one and its information has been gathered through a library method. The results of the research showed that the time behavior of the daily price of stock companies and indices is not random in the stock exchange, but this non-random process has many complications. When using neural networks for prediction of the design of the neural network model, it is needed to use the network with an appropriate number of layers and intermediate neurons.