KLASIFIKASI PERUBAHAN HARGA OBLIGASI KORPORASI DI INDONESIA MENGGUNAKAN METODE NAIVE BAYES CLASSIFICATION

Published: 29 Apr 2016.
Open Access

Citation Format:
Abstract
Bond is a medium-long term debt securities which can be sold and contains a pledge from the issuer to pay interest for a certain period and repayment of the principal debt at a specified time to the bonds buyer. Bonds price changes any time, it could be beneficial or give disadvantage to investors. Investors should know the best conditions to buy bonds on a discount, or sell them at a premium price. By classify the changing of bonds price, it could help investors to gain optimum return. One method is Naive Bayes classification. In theory, It has the minimum error rate in comparison to all other classifiers. Bayes is a simple probabilistic-based prediction technique which based on the application of Bayes theorem with strong independence assumptions. Before classifying, preprocessing data is required as a stage feature selection. In this case, the Mann Whitney test can be done to choose the independent features of each class. Validation technique in use is k-fold cross validation. Based on analysis, we gained average accuracy at 78,52% and 21,8% error. With high accuracy and quite low error, it means that the Naïve Bayes method works quite well on  classifying the corporate bonds price changes in Indonesia. Keywords: bonds, classification, k-fold cross validation, Naive Bayes
Keywords: bonds; classification; k-fold cross validation; Naive Bayes

Article Metrics: