Open AccessPredicting factors affecting the future stock price crash risk based on support vector machine
Author(s) -
Mohammad Reza Razdar,
Ali Mohammad Zahmatkesh,
Sanaz Khaleh Oghlizadeh
Publication year - 2019
Publication title -
journal of management and accounting studies
Language(s) - English
Resource type - Journals
ISSN - 2693-8448
DOI - 10.24200/jmas.vol5iss03pp7-14
Subject(s) - support vector machine , stock exchange , shareholder , computer science , sample (material) , crash , mean absolute percentage error , stock price , stock (firearms) , machine learning , econometrics , artificial intelligence , business , finance , engineering , artificial neural network , economics , series (stratigraphy) , corporate governance , mechanical engineering , paleontology , chemistry , chromatography , biology , programming language
The prediction of stock price crash risk is an important and widely studied topic in both accounting and finance, since crash risk has a significant impact on shareholders, creditors, managers, investors, and regulators. The aim of this research is to analyse Predicting factors affecting the future stock price crash risk based on support vector machine.Methodology:In this research we study the data of 99 companies listed on the Tehran Stock Exchange (TSE) from 2011 to 2016.And since the Mean Absolute Error in the Testing Sample is less than Training, then the model estimation is possible using the support vector machine method.Results:The results shows the summary of support vector machine Model. The results indicate support vector machine includes 3 input layers, and 1 output layer. Conclusion:The method used in this study support vector machine is a kind of RBF. The Mean Absolute Error for the Training sample is 0.078 and for the Testing example is 0.065. And since the Mean Absolute Error in the Testing Sample is less than Training, then the model estimation is possible using the support vector machine method.