Customization of User Interactions with Data-intensive Systems by Modeling of User Communication Style, Temperament, and Preferences by Anne Y.A. Chen

Current web-search engines provide query results with relatively high precision and recall, but user satisfaction is often low. Precision and recall are two traditional measurements of data accuracy but not user satisfaction and preferences. Therefore, user characteristics and preferences have become important in information retrieval. Modeling users' temperaments provides a base for an information recommendation service. We incorporate human temperaments into the filtering process via Keirsey's temperament theory, probability theory, and statistical reasoning. The system providing recommendation services searches information taking into account user temperament (communication style). According to Keirsey's theory, there are four types of temperaments which identify users' inner personalities. These are SJ (Sensing and Judging), SP (Sensing and Perceiving), NT (iNtuition and Thinking) and NF (Intuition and Feeling). We find correlations between user temperaments and user preferences, which provides a general foundation for our content-based filtering recommendation system.

Although many recommendation systems are designed to provide personalized query results to match user preferences in order to increase user satisfaction, none of these systems was designed to interpret the semantic meaning of user queries and match it with semantic information of the data. The goal of our research is to test the hypothesis that a semantics-based system incorporating ontologies can provide personalized query results that match user preferences and increase user satisfaction. A system is being developed to examine the effects of customized information retrieval. This semantics-based system consists of six components: query processor, query generation, user information collector, user profile manager, result refinement, and ontology management. The information in ontologies was incorporated into the information process functioning of every component in the system. The user information collector and user profile manager compile user information and analyze that information based on the ontologies. Information in a user profile is applied in order to provide customized result presentations that match user preferences. The semantics-based customization can be associated with online search engines, online shopping sites, and other web service information providers.