文章基本信息

标题：BACKPROPAGATION NEURAL NETWORK AND CORRELATION-BASED FEATURE SELECTION FOR EARNING RESPONSE COEFFICIENT PREDICTION
本地全文：下载
作者：ABDUL SYUKUR ; CATUR SUPRIYANTO
期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2014
卷号：67
期号：1
出版社：Journal of Theoretical and Applied
摘要：This paper evaluates the prediction of Earning Response Coefficient (ERC) through data mining. The collected data included 10 variables which are earning persistance, firm size, systematic risk, earning growth, earnings predictability, operating leverage, financial leverage, barrier to entry, transaction gains (losses) and ERC as a target prediction. Backpropagation Neural Network (BPNN) and correlation feature selection are applied in order to predict ERC which is trained and tested using 10-fold validation. Samples used in this study are 241 firms listed in the Jakarta Stock Exchange (JSE) from 2000-2002. The results of experiments achieve two main finding: BPNN and correlation feature selection perform well to predict ERC and our prediction model is capable to select the relevant attribute for the prediction.
关键词：Earning Response Coefficient; Backpropagation Neural Network; Prediction.