KLASTERISASI HARGA SAHAM DAN KOMODITAS MENGGUNAKAN METODE HYBRID KLASTERISASI
| Main Author: | Santoso, Halim Budi |
|---|---|
| Format: | Thesis NonPeerReviewed Book |
| Terbitan: |
, 2013
|
| Subjects: | |
| Online Access: |
http://e-journal.uajy.ac.id/4424/1/0MTF01830.pdf http://e-journal.uajy.ac.id/4424/2/1MTF01830.pdf http://e-journal.uajy.ac.id/4424/3/2MTF01830.pdf http://e-journal.uajy.ac.id/4424/4/3MTF01830.pdf http://e-journal.uajy.ac.id/4424/5/4MTF01830.pdf http://e-journal.uajy.ac.id/4424/6/5MTF01830.pdf http://e-journal.uajy.ac.id/4424/ |
Daftar Isi:
- Stock price volatility is one of the investment risks. Investor should have updated information. Based on those information, Investor will decide whether want to buy, sell, or hold the stock. Stock price volatility is influenced by some factors related with the global economy. One of the them is commodity price. If the commodity price were to drop sharply, is there any impact to the specific company. Based on this question, this study is conducted. This study is conducted to test the K-Means Clustering, Principal Component Analysis, and Backpropagation Neural Network for stock price and commodity price. K-Means Clustering is conducted to group the similar movement of the stock price and commodity price. After the cluster is established, Principal Component Analysis is implemented for every cluster. Besides, Artificial Neural Network is implemented to every cluster using Backpropagation Neural Network. There will be comparation study about the performace of the Artificial Neural Network for compact cluster and uncompact cluster. Implementation of the K-Means Clustering will make ten cluster of commodity price and the effect to the spesific sector. This study also found that Principal Component Analysis can reduce the dimension of every cluster. Principal Component Analysis also can improve the performace by reducing the Mean Squared Error of every network for every cluster.