Abstract—This paper focuses on the text mining approach of the gold prices volatility prediction model from the textual of economic indicators news articles. The model is designed and developed to analyze how the news articles influence gold price volatility. The selected reliable source of news articles is provided by FXStreet which offers several economic indicators such as Economic Activity, Markit Manufacturing PMI, Bill Auction, Building Permits, ISM Manufacturing Index, Redbook index, Retail Sales, Durable Goods Orders, etc. The data will be used to build text classifiers and news group affecting volatility price of gold. According to the fundamental of data mining process, each news article is firstly transformed in to feature by TF-IDF method. Then, the comparative experiment is set up to measure the accuracy of combination of two attributes weighting approaches, which are Support Vector Machine (SVM) and Chi-Squared Statistic, and three classification algorithms, which are the k-Nearest Neighbour, SVM and Naive Bayes. The results show that the SVM method is the most superior to other methods in both attributes weighting and classifier viewpoint.
Index Terms—text mining, economic indicators news, gold price, volatility prediction
Cite: Chanwit Onsumran, Sotarat Thammaboosadee, and Supaporn Kiattisin, "Gold Price Volatility Prediction by Text Mining in Economic Indicators News," Vol. 6, No. 4, pp. 243-247, November, 2015. doi: 10.12720/jait.6.4.243-247
Copyright © 2013-2021. Journal of Advances in Information Technology. All Rights Reserved
This work is licensed under the Creative Commons Attribution License (CC BY-NC-ND 4.0)