Q65 : Developing a feature selection method for Iranian stock prediction
Thesis > Central Library of Shahrood University > Computer Engineering > MSc > 2015
Authors:
Abstarct: Analysis of financial market always attract a lot of attention from investors and researchers to have. The price of the stock market is complex and affected by many factors. So find the major factors affecting the stock market is very important and very difficult to predict the other hand, more recent articles researchers trying to predict the market with a series of features, usually with the aim of predicting the market price is, the results of it is unusable in the real market, so we decided to look from a different angle and applied research to the market to help traders do stock market, in which we seek to predict the future rather than the predicted price market that is divided into two categories descending and ascending, in this study we aim to test different parameters and find the influential factors that can predict future market uptrend or downtrend help us, things that we in this thesis we examined markers (indicators), which traders exploit them when predicting market trends. In order to predict the clustering support vector machine was used to identify factors affecting the rapper backward feature selection method was used to remove the characteristics of ineffective health and the properties remain effective. By implementing these algorithms on Shahrood cement stocks and Pars Oil could be the indicators most effective with respect to any stocks acquired time series.
Keywords:
#Support vector machine #predict stock trends #feature selection #indicator Link
Keeping place: Central Library of Shahrood University
Visitor:
Keeping place: Central Library of Shahrood University
Visitor: