Earticle

초록

영어

In view of traditional collaborative filtering recommendation does not take into account differences of dimension of user vector and value of evaluation, this paper proposed a collaborative filtering recommendation model based on normalization method. Before calculating the user’s or item’s similarity, the value of evaluation will be normalized to a range of specifications. Then the similarity of user vector will be calculated, and predictions and recommend will be made. The experimental results show that this model could accurately find similar neighbor users or items, and performances of prediction and recommendation have been largely improved.

목차

Abstract
 1. Introduction
 2. Principle of Traditional Collaborative Filtering Recommendation
  2.1. Data Definition
  2.2. Similarity Calculation
  2.3. Item Rating Prediction and Recommendation
 3. Collaborative Filtering Recommendation Based on Normalized Method
  3.1. Deficiency of Traditional Collaborative Filtering Algorithm
  3.2. Similarity Measurement Method Based on Normalization Method
  3.3. Item Rating Prediction and Recommendation Based on Normalization Method
 4. Experiment and Analysis
  4.1. Experimental Data Set
  4.2. Evaluation Criteria
  4.3. Experimental Scheme
 5. Conclusion
 References

키워드

저자

• Gao Yan [ Information Engineering School, Yulin University, Yulin, Shaanxi, China ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Grid and Distributed Computing

• 간기

  격월간

• pISSN

  2005-4262

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605