微型电脑应用

为获得更加理想的电子商务推荐结果，提出一种基于协同过滤的电子商务智能推荐方法。该方法收集电子商务用户相关信息，并对信息进行预处理，计算电子商务用户对项目评分，构建电子商务用户评分矩阵，采用余弦算法根据用户评分矩阵计算用户之间的相性度，基于用户相似度进行电子商务智能推荐。为了与其他方法进行比较，开展仿真实验。实验结果表明，提出的方法较好地解决了当前电子商务推荐方法存在的缺陷，提高了电子商务推荐精度，减少了电子商务推荐的错误率，具有明显的优越性。

In order to obtain more ideal e-commerce recommendation results, an e-commerce intelligent recommendation method is proposed based on collaborative filtering. This method collects the relevant information of e-commerce users, preprocesses the information, calculates the item score of e-commerce users and the user score matrix, finally uses cosine angle algorithm to calculate the phase degree among users according to the user score matrix, and makes e-commerce intelligent recommendation based on the user similarity degree. A simulation experiment of e-commerce recommendation is carried out for comparing with other methods. The results show that this method solves the defects of current e-commerce recommendation methods, improves the accuracy of e-commerce recommendation, reduces the error rate of e-commerce recommendation, and has obvious advantages.