Abstract: In e-commerce times, it is important for an e-commerce website whether it can provide better personality and individuality services for customers. An adaptive recommend framework was presented and accuracy of recommendation was improved through combining some basic recommendation algorithms. Average method and grouping adjustment method were given to solve combination problems of these basic algorithms. The experiment shows that the two methods have well recommend results. The framework can avoid defects of single algorithm, and it is more suitable for recommend application of e-commerce.

Key words: Recommend system, Collaborative filtering, Adaptive

摘要： 在电子商务时代，提供更加人性化、个性化的服务是一个电子商务网站生存的关键。提出一种自适应推荐框架，通过组合各种原子推荐算法来提高推荐的准确性；提出采用平均法和分组调整法两种机制来处理各原子算法的结合问题。实验表明，这两种方法都取得了良好的推荐效果。该框架能够有效避免单一推荐算法的缺陷，更好地适应电子商务推荐应用。

关键词: 推荐系统, 协同过滤, 自适应