互联网目前已经进入了信息爆炸的时代，面对海量的信息，即使通过技术成熟的搜索引擎，用户往往还是很难快速有效地获取到自己想要的信息。为了更有效地解决这一难题，推荐系统应运而生，成为了配合搜索引擎的重要手段。在推荐系统中，用户不需要通过使用关键词或提出明确的目标需求，即可寻找到自己潜在的目标信息。推荐系统这一技术目前在电子商务、新闻及视频推荐等领域都取得了十分广泛的应用并产生了良好的效果。 协同过滤推荐算法是目前推荐系统领域中十分常用的方法，余弦相似度和Pearson相关系数是目前协同过滤推荐算法中计算相似度的两种常用算法。为提高协同过滤推荐算法的准确性，对相似度计算问题进行了研究，针对目前常用的余弦相似度和Pearson相关系数这两种相似度计算方法的不足，通过设计和引入调节因子，分别考虑用户在评分习惯和项目选择上的差异性，来对这两种传统的相似度算法进行优化和改进。另外考虑到用户的偏好往往与项目所具有的属性有关，设计出衡量用户对属性偏好的参数，通过加权的方式，与改进后的相似度算法进行融合，提出了一种融合用户评分习惯、项目选择差异及属性偏好的协同过滤推荐算法。通过在MovieLens数据集上进行实验的结果表明，研究提出的改进算法相比于传统算法更为精确，平均绝对误差和均方根误差有了明显的降低。 最后，将提出的融合用户对项目和属性偏好的协同过滤推荐算法在图书馆的座位推荐这一场景进行了应用方案设计，并以北京师范大学图书馆为例进行了实践应用，将理论与实践进行了结合。

The Internet has entered the era of information explosion, when dealing with vast amounts of information, even with the mature technology of search engine, users are still difficult to obtain the information they want quickly and efficiently. In order to solve this problem, recommendation system emerged as an important tool to cooperate with the search engine. In recommendation system, users do not need to use keywords or give specific requirements and they will find information of their potential target. Recommendation system has been widely used in e-commerce, news and video recommendation, and it has worked well. Collaborative filtering algorithm is one of the most successful and useful technologies in recommendation systems. Cosine similarity and Pearson correlation coefficient are two of the most widely used traditional algorithms to calculate the user similarity in collaborative filtering algorithm. In order to reduce the error, an improved collaborative filtering recommendation algorithm is proposed in view of the disadvantages of the two traditional similarity algorithms. The two traditional algorithms are improved by importing two parameters, one of them is proposed for considering the custom of user’s rating, and the other is imported to measure the difference of items chosen by users. Besides, usually people like items for some attributes, and a parameter is designed to measure it. The new algorithm is constructed by the improved traditional algorithm and user’s preference for attributes. The results of experiment based on MovieLens dataset show lower mean absolute error (MAE) and root mean square error (RMSE), and prove better performance by using the two parameters to improve the traditional similarity algorithms. Finally, the new collaborative filtering algorithm based on user’s preference for items and attributes is applied in seat recommendation, and the specific application solution is given. The practice of the solution is applied in the library of Beijing Normal University, which achieve the combination of theory and practice.