Customization Using Support Vector Machines for Information Retrieval

This paper proposes an HCI approach to supporting interactive document retrieval in unorganized open information space. On the assumption that taxonomical thought are one of the most important and skilled operations for us when we organize or store information, we may discuss that clustering or classification techniques are applied to interactive interface. The proposed system, therefore, provides two-dimensional space to visualize categories of document icons for interactive interface. The features of the proposed approach are (1) visualization of document categories for interaction, (2) initialization of categories by hierarchical clustering method, (3) customization of categories by support vector machine techniques, (4) user-depend additional attributes for individual implicit, and (5) cognitive aspects integration of these techniques for IR interface. We made a preliminary experiment on text categorization for customization, using two test collections to estimate practicability.