CAJ | 학술논문

对个性化推荐系统算法进行改进,首先,不仅考虑用户所选的商品,而且考虑用户的打分,从而将资源分配法扩展为含权资源分配;其次,考虑用户的同有相似性.把这两方面相结合,发展了新的算法.数值试验表明,改进后的方法显著提高了推荐的精度和个性化程度.
대개성화추천계통산법진행개진,수선,불부고필용호소선적상품,이차고필용호적타분,종이장자원분배법확전위함권자원분배;기차,고필용호적동유상사성.파저량방면상결합,발전료신적산법.수치시험표명,개진후적방법현저제고료추천적정도화개성화정도.
This paper presents an improved algorithm for personal recommendation. It takes not only the merchandise selected by users into account, but also the ratings given by the users. Then,it considers the inherent similarities between the users. Combining these two factors together, a new algorithm is developed. The numerical simulation results show that the presented algorithm can obviously enhance the accuracy and personality of the recommendation.