Knowledge-Domain Semantic Searching and Recommendation Based on Improved Ant Colony Algorithm

To obtain accurate search results and advocate the use of human effort in discovering knowledge, we propose a method based on Ant Colony Algorithm (ACA). The proposed method simulates the behavior of ants searching for food. Specific features such as the behavior of ants searching for food, their established search paths, and the ant "neighborhood" profile are investigated. The investigation results reveal that the behavior of people searching for useful information resembles that of ants searching for food. We also use semantic annotation and the decreasing matrix dimension approach to accelerate the food searching process and shorten the distance between the query starting points and the ultimate answers. A user behavior model is constructed based on personal and domain ontologies. Experimental evaluation with the enhanced ACA has two parts: (1) estimating the efficiency of information retrieval with user interests considered and (2) identifying how to weigh usage and rate user data during recommendation.