Class Association Rules for Profiling Outlier Stocks


Ng, Keng Hoong and Khor, Kok Chin and Tong, Gee Kok (2017) Class Association Rules for Profiling Outlier Stocks. International Journal of Advances in Soft Computing and its Application, 9 (3). pp. 114-131. ISSN 2074-8523

[img] Text
Restricted to Registered users only

Download (659kB)


Finding a stock with superior financial performance demands not only abundance of time, but a lot of financial knowledge from retail investors. Consequently, they always end up with empty handed. This research aims to assist them to “recognize” this type of stock in a fast manner, despite they are not financially savvy. In this study, we started with identifying outliers in a pool of construction stocks. Then, these outliers were manually classified into two classes, i.e. outstanding or poor outliers. Class association rule mining was performed to these classes to generate sets of association rules, which were used to profile each outlier class. Investors may use the rules of the profiles to pick potential outstanding stocks or avoid poor performance stocks.

Item Type: Article
Uncontrolled Keywords: construction stocks, financial ratios, local outlier factor, data discretization, association rules mining
Subjects: Q Science > QA Mathematics > QA299.6-433 Analysis
Divisions: Faculty of Computing and Informatics (FCI)
Depositing User: Ms Nurul Iqtiani Ahmad
Date Deposited: 04 Dec 2017 14:14
Last Modified: 04 Dec 2017 14:14


Downloads per month over past year

View ItemEdit (login required)