Mining User Behavioral Rules from Smartphone Data through Association Analysis
This work addresses the problem of inefficient decision-making due to redundant rules for researchers and practitioners in mobile data analysis, but it is incremental as it builds on existing association rule learning techniques.
The paper tackles the problem of mining behavioral association rules from smartphone data by addressing the issue of redundant rules, proposing an approach that identifies redundancy and extracts a concise set of non-redundant rules, with effectiveness examined on real mobile phone datasets.
The increasing popularity of smart mobile phones and their powerful sensing capabilities have enabled the collection of rich contextual information and mobile phone usage records through the device logs. This paper formulates the problem of mining behavioral association rules of individual mobile phone users utilizing their smartphone data. Association rule learning is the most popular technique to discover rules utilizing large datasets. However, it is well-known that a large proportion of association rules generated are redundant. This redundant production makes not only the rule-set unnecessarily large but also makes the decision making process more complex and ineffective. In this paper, we propose an approach that effectively identifies the redundancy in associations and extracts a concise set of behavioral association rules that are non-redundant. The effectiveness of the proposed approach is examined by considering the real mobile phone datasets of individual users.