基于时间和空间的推荐方法研究以及应用

发布时间：2018-08-28 11:07

【摘要】：随着计算机网络通信技术和Internet的飞速发展,信息过载是互联网面临的主要挑战。随着电子商务的发展,商家提供的商品种类和数量急剧增长.具有明确需求的用户可通过搜索查找想购买的商品。然而,用户需求通常具有不确定性和模糊性。随着人们生活水平的提高,汽车的使用也越来越普遍。庞大的汽车使用人群,使得汽车的维修需求量日益增大。如何实时并且准确地向汽车维修用户提供最专业的服务,为他们提供汽车维修店的完整的、具体的信息是汽车维修商和车主共同面临的问题。个性化推荐系统是解决这一问题的有效手段。本文作者参加了阳扬车诊APP设计与开发的部分工作,根据项目的需要,针对汽车的流动性和汽车维修数据以及动态性特点,从时间和空间的角度研究了对汽车维修的个性化推荐方法,并设计了相应的论文推荐系统,主要研究工作和成果如下:(1)提出了基于时间的推荐算法。将用户-项目的评分矩阵中加入时间这个因素,将久远的评分影响力降低,最近的评分影响力提高。对用户之间的相似度考虑时间因子,找出用户对其尚未表达的项目的潜在兴趣度,用协同过滤的算法来解决推荐问题。实验结果说明该算法的推荐结果比较其他相类似的推荐算法具有较高的精度。(2)提出了基于地理信息的推荐算法。将相似度和地理信息相结合进行推荐,不仅计算用户与用户之间的地理相似度,还计算用户与项目之间的地理相似度。为了推荐的效果更好,该算法还将地理距离过大的用户与项目之间的相似度默认为零。实验结果说明,该算法的推荐结果的准确性有所提升。(3)根据阳扬车诊APP的需要,设计与开发了汽车维修推荐系统,在该系统中应用了本文所提出的基于时间的推荐算法和基于地理信息的推荐算法。进行了对该系统的需求分析,并且将总体框架设计成多模块层次结构,实现了向用户推荐维修点和回答的推荐功能。
[Abstract]:With the rapid development of computer network communication technology and Internet, information overload is the main challenge facing the Internet. With the development of e-commerce, the types and quantity of goods provided by merchants are increasing rapidly. Users with specific needs can search for items they want to buy. However, user requirements are usually uncertain and fuzzy. With the improvement of people's living standards, the use of cars is becoming more and more common. The large number of automobile users makes the demand for vehicle maintenance increasing day by day. How to provide the most professional service to the automobile maintenance users in real time and accurately, and how to provide them with the complete information of the automobile repair shop is a common problem faced by the vehicle maintenance and vehicle owners. Personalized recommendation system is an effective method to solve this problem. The author takes part in the design and development of Yangyang vehicle diagnosis APP. According to the need of the project, according to the vehicle mobility and vehicle maintenance data and dynamic characteristics, The personalized recommendation method for vehicle maintenance is studied from the angle of time and space, and the corresponding thesis recommendation system is designed. The main research work and results are as follows: (1) A time-based recommendation algorithm is proposed. Adding the time factor to the user-project rating matrix reduces the impact of a long-term rating and increases the impact of the most recent rating. Considering the time factor to the similarity between users, we find out the potential interest of users in the items they have not yet expressed, and use collaborative filtering algorithm to solve the recommendation problem. The experimental results show that the proposed algorithm has higher accuracy than other similar recommendation algorithms. (2) A recommendation algorithm based on geographic information is proposed. By combining similarity with geographic information, not only the geographic similarity between users and users is calculated, but also the geographic similarity between users and items is calculated. In order to make the recommendation more effective, the similarity between the user and the project is zero by default. The experimental results show that the accuracy of the proposed algorithm is improved. (3) according to the need of Yangyang vehicle diagnosis APP, a vehicle maintenance recommendation system is designed and developed. In this system, the time-based recommendation algorithm and the geographic information-based recommendation algorithm are applied. The requirement analysis of the system is carried out, and the overall framework is designed into a multi-module hierarchical structure, and the function of recommending maintenance points and answers to users is realized.
【学位授予单位】：扬州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3

【参考文献】