A Novel Approach for Frame-Based Knowledge Similarity

This study proposes the RF-RS-RV structure for frame-based knowledge and integrates methods of conditional probability, vector matrices, and rule-based inference to establish a frame-based knowledge similarity algorithm. This algorithm can accurately calculate frame-based knowledge similarity matrices and determine the relationship among knowledge items, this relationship can function as the knowledge source for conflicts treatments and increase the added-value of the knowledge, so wrong decisions be avoided.