PREDIKSI HARGA SAHAM MENGGUNAKAN SUPPORT VECTOR REGRESSION DENGAN ALGORITMA GRID SEARCH

*Hasbi Yasin - 
Alan Prahutama - 
Tiani Wahyu Utami - 
Published: 30 Jun 2014.
Open Access
Citation Format:
Article Info
Section: Articles
Language: EN
Full Text:
Statistics: 2223 1999
Abstract

The stock market has become a popular investment channel in recent years because of the low return rates of other investment. The stock price prediction is in the interest of both private and institution investors. Accurate forecasting of stock prices is an appealing yet difficult activity in the business world. Therefore, stock prices forecasting is regarded as one of the most challenging topics in business. The forecasting techniques used in the literature can be classified into two categories: linear models and non linear models.  One of forecasting techniques in nonlinear models is support vector regression (SVR). Basically, SVR adopts the structural risk minimization principle to estimate a function by minimizing an upper bound of the generalization. The optimal parameters of SVR can be use Grid Search Algorithm method. Concept of this method is using cross validation (CV). In this paper, the SVR model use linear kernel function. The accurate prediction of stock price, in telecommunication, is 92.47% for training data and 83.39% for testing data.

 

Keywords: Stock price, SVR, Grid Search, Linear kernel function.

Article Metrics: