유전자 알고리즘을 이용한 개인화된 정보 필터링 = Personalized Information Filtering using Genetic Algorithm

다국어 초록 (Multilingual Abstract)

In this paper, the Genetic Algorithm for searching the optimized weighting values for the importance of words in the user profile and the importance of the position of words in searched documents are proposed. Genetic operators such as selection, crossover and mutation are developed and fitness function is derived. The optimized values are calculated using proposed Genetic Algorithm. The optimized weighting values are applied to the personalized information filtering problem. The similarity of searched documents are calculated using weighting values for the importance of words or that of the position of words. The priority of documents are rearranged by their similarity. The difference of priorities based on weighting values and user's subjective decision of user are evaluated and the difference is seed as fitness of each chromosome in Genetic Algorithm. The chromosome that has smallest fitness values is chosen as optimized weighting value. It has been shown by simulations such that the proposed schema can be applied to the personalized information filtering.

목차 (Table of Contents)

• 목차 = 3
• Ⅰ. 서론 = 1
• Ⅱ. 관련 연구 = 3
• 1. 웹 문서에 대한 정보 검색 = 3
• 2. 검색 모델 = 5
• 3. 정보필터링 = 6
• Ⅲ. 유전자 알고리즘을 이용한 개인화된 정보 필터링 = 8
• 1. 유전자 알고리즘 = 8
• 2. 개인화된 정보 필터링을 위한 사용자 프로파일 단어 가중치 학습을 위한 유전 알고리즘 = 10
• 2.1 유전자 알고리즘의 염색체 구조 = 12
• 2.2 유전자 알고리즘 연산자(Select, Crossover, Mutation) = 12
• 2.3 적합도(Fitness) = 18
• Ⅳ. 실험 및 고찰 = 20
• Ⅴ. 결론 = 44
• 참고문헌 = 45
• Abstract = 48