Investigating the Prices Manipulation in the Tehran Stock Exchange by Using the SVM Model

• Main Authors: Mir Feyz Falah Shams, Mahdi Rashnoo
• Format: Article
• Language: fas
• Published: University of Tehran 2012-07-01
• Series: تحقیقات مالی
• Subjects: price manipulation; duration dependence test; support vector machine
• Online Access: https://jfr.ut.ac.ir/article_36633_c596a39bc83bb0a2e6004b9fa02c8a41.pdf

Phenomenon of price manipulation is one of the factors which have caused mistrust of the investors to the stock market and inhibits its growth and prosperity. Entering the shareholders into the stock market, on one hand leads to increase in the general level of revenues and on the other hand causes inexpensive financing for companies. In this research, at first by using duration dependence test and among the 379 companies, 95 cases were identified as the manipulated companies. Then prediction accuracy of SVM model on prices manipulation in the stock market was examined. SVM model is one of the models which is used for classifying and separating the groups which their under examining data should be linear. It was tried to resolve the problem of not being linear data by using PCA. The research results show that the model correctly predicted to the extent of 81percent of manipulations