Discovering generalized association rules from Twitter
Intelligent Data Analysis
| Hi-index | 0.00 |
Context-aware systems acquire and exploit information on the user context to tailor services to a particular user, place, time, and/or event. Hence, they allow service providers to adapt their services to actual user needs, by offering personalized services depending on the current user context. Service providers are usually interested in profiling users both to increase client satisfaction and to broaden the set of offered services. Novel and efficient techniques are needed to tailor service supply to the user (or the user category) and to the situation in which he/she is involved. This paper presents the CAS-Mine framework to efficiently discover relevant relationships between user context data and currently asked services for both user and service profiling. CAS-Mine efficiently extracts generalized association rules, which provide a high-level abstraction of both user habits and service characteristics depending on the context. A lazy (analyst-provided) taxonomy evaluation performed on different attributes (e.g., a geographic hierarchy on spatial coordinates, a classification of provided services) drives the rule generalization process. Extracted rules are classified into groups according to their semantic meaning and ranked by means of quality indices, thus allowing a domain expert to focus on the most relevant patterns. Experiments performed on three context-aware datasets, obtained by logging user requests and context information for three real applications, show the effectiveness and the efficiency of the CAS-Mine framework in mining different valuable types of correlations between user habits, context information, and provided services.