企业知识个性化推荐方法研究与应用

发布时间：2018-03-10 17:51

本文选题：推荐算法　切入点：知识推荐　出处：《齐鲁工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：当今社会环境下,优秀的企业积累了很多历史数据,这些历史数据包含丰富的经验和知识。优秀企业会特别注重这些历史数据,因为历史数据中往往包含着某些重要信息和行业发展趋势,我们把包含重要信息的历史数据称之为企业知识,这些也就是企业的财富。随着互联网技术的发展与普及,知识传播的速度越来越快,企业知识呈现指数级增长,员工在单位时间内获取到合适的知识的效率越来越低。因为知识库存储的知识越来越多,知识库的利用率也就慢慢下降。到目前为止,有效解决此问题的技术可以分为以推荐系统为代表的信息过滤技术和以搜索引擎为代表的信息检索技术。本文主要研究的是前者。随着电子商务的发展和个性化推荐系统的问世,有许多推荐算法被提出,而且在特定领域都发挥着不可替代的作用,比如说应用在电子商务领域中的协同过滤算法。虽然此算法被广泛应用,但是其仍然存在着诸如数据稀疏问题、冷启动等问题。随着企业知识库的飞速发展,企业知识的推荐也逐渐成为最近的研究热门。如何有效的利用相关技术来改进推荐算法的性能和提高企业知识推荐质量也日益被科研人员广泛研究。针对现有的问题,本文在以下两个方面展开研究。第一,关于马尔可夫预测模型的研究。本文针对用户-项目之间的联系,认为用户本次查看项目行为和下次查看行为是有强联系的,基于这个设想提出一种基于马尔可夫预测模型的协同过滤推荐算法。实验结果显示,此算法具有较好的推荐效果。第二,关于粗糙集理论与企业知识的相关研究。针对员工的需要自动为其推荐恰当的知识,可以提高员工的工作效率和员工与公司的粘度,同时可以提高企业知识的应用,对于员工和企业都是有利的。把员工需求用粗糙集理论表示出来,然后把知识库中的知识与其模型相匹配,得到推荐结果。此模型可以充分利用知识库中的知识,有利于企业的发展与进步。
[Abstract]:In today's social environment, excellent enterprises have accumulated a lot of historical data. These historical data contain rich experience and knowledge. Good enterprises pay special attention to these historical data. Because historical data often contain some important information and industry trends, we call historical data containing important information as corporate knowledge, which is the wealth of enterprises. With the development and popularization of Internet technology, The speed of knowledge transmission is getting faster and faster, the enterprise knowledge is increasing exponentially, and the efficiency of obtaining the appropriate knowledge in a unit of time is becoming increasingly low, because the knowledge base stores more and more knowledge, The utilization of the knowledge base is declining. So far, The effective technology to solve this problem can be divided into the information filtering technology represented by recommendation system and the information retrieval technology represented by search engine. The former is mainly studied in this paper. With the development and individuation of electronic commerce. The advent of the recommendation system, Many recommended algorithms have been proposed and play an irreplaceable role in specific areas, such as collaborative filtering algorithms used in electronic commerce. Although this algorithm is widely used, However, there are still some problems such as sparse data, cold start, etc. With the rapid development of enterprise knowledge base, How to improve the performance of recommendation algorithm and improve the quality of enterprise knowledge recommendation has been widely studied by researchers. In this paper, the following two aspects of the study. First, the study of Markov prediction model. According to the relationship between the user and the project, this paper thinks that there is a strong relationship between the behavior of the user viewing the project and the next viewing behavior. Based on this assumption, a collaborative filtering recommendation algorithm based on Markov prediction model is proposed. The experimental results show that the algorithm has a good recommendation effect. Second, The relevant research on rough set theory and enterprise knowledge. According to the needs of employees, it can automatically recommend appropriate knowledge, which can improve the efficiency of employees and the viscosity between employees and companies, at the same time, it can improve the application of enterprise knowledge. It is beneficial to both employees and enterprises. The employee needs are expressed by rough set theory, then the knowledge in the knowledge base is matched with its model, and the recommended results are obtained. This model can make full use of the knowledge in the knowledge base. It is conducive to the development and progress of enterprises.
【学位授予单位】：齐鲁工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.3

【相似文献】