An adaptive soft classification model: Content-based similarity queries and beyond

Providing a customized result set based upon user preference is the ultimate objective of many information-locating systems. There are two main challenges in meeting this objective: First, there is a gap between the physical characteristics and the semantic meaning of the multidimensional data. Second, different people may have different perceptions on the same set of data. To address both these challenges, we propose a model that brings together the advantages of content-based querying and collaborative filtering techniques. Our model first conceptualizes the information locating work as the task of "soft classifying" items into classes. These classes can overlap, and their members are different for different users. The soft classification is hence performed for each and every item feature including both physical and semantic features.; Subsequently, each item will be ranked based on the weighted aggregation of its classification memberships. The weights are user dependent and hence different users would obtain different result sets for the same query. Our model employs a fuzzy-logic based aggregation function for ranking items. We show that in addition to some performance benefits, fuzzy aggregation is less sensitive to noise and can support disjunctive queries as compared to the weighted average method used by many information retrieval systems.; Finally, since our technique heavily relies on user dependent weights (i.e., user profiles) for the aggregation task, we utilize the users' relevance feedback to improve the profiles using genetic algorithms (GA). Our learning mechanism requires less user interaction and results in faster convergence to the user's preferences as compared to other learning techniques.; We utilize our customization technique in three different applications: (1) a content-based image retrieval system, (2) an e-commerce recommendation system, and (3) a neuroscience knowledge management system. Our experimental results demonstrate while the complexity of the soft classification system is low and remains constant as the number of users and/or items grow, its accuracy surpasses those of many information retrieval systems in most cases. The experimental results also indicate that the retrieval accuracy is significantly increased by using the GA-based learning mechanism.