面向产学合作的高校推荐研究与应用

发布时间：2018-01-12 04:04

本文关键词：面向产学合作的高校推荐研究与应用　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：在"大众创业,万众创新"的背景下,创新驱动发展战略的部署和落实成为我国政府的工作重点,其中,在我国提高自主创新能力的道路上,引导和支持创新要素向校企汇集并促进科技成果转化为生产力的产学合作起着不可替代的作用。随着产学合作相关工作的不断推进,也发现了其中不少的问题:产学合作的深度有待加深;高校的科技成果与真实的市场需求无法有效衔接从而影响产学合作的质量;由于缺乏相应的渠道和信息获取平台,容易造成校企双方信息的不对称,一般作为需求方的企业选择合作高校的方式单一,多是通过学校名气或熟人介绍,在一定程度上会导致高校相关能力无法与企业的需求进行匹配,使得知名度不太高但在相关领域专业能力较强的高校失去合作的机会,同时企业资源也无法得到合理配置。随着互联网的发展,信息的超载问题使得信息的筛选和过滤变得十分重要,如何利用先进的大数据技术,通过海量数据的分析为产学两方解决目前的信息不对称问题,并选择合适的产学合作伙伴,从而实现资源的更合理地配置,提高产学合作质量和效率成为本研究的重点。本文面向产学合作,通过高校产学合作现状的分析和相关文献的研究,梳理了高校产学合作能力的关键指标,为高校推荐模型的构建奠定基础。从企业需求出发,对需求内容本身进行文本预处理、特征提取和关键词分类,使企业需求能够被较准确地分到某一领域或类别中。再结合产学合作的特点,站在企业选择合作伙伴的角度,在"高校基础科研能力"、"高校成果转化能力"、"高校产学合作背景"这三个一类关键指标的基础上继续细化和丰富指标,结合专家意见,采用5级标度赋值法,通过最小二乘法进行回归微调,最后确定了各级指标的权重值。基于企业需求类别信息,利用网络爬虫相关技术,引入asynchttpclient包和htmlparser包,获取高校的一部分与类别相匹配的产学信息,包括论文及科研项目,共同构建出基于企业需求的高校推荐算法模型,接着通过一个算例在一定程度上对算法进行了验证。最后采用面向对象的方法,基于UML对面向产学合作的高校推荐系统进行了分析建模与设计,并借助AxureRP原型工具,设计了系统的界面,完成了 Web页面的展示和交互,使得推荐模型进一步得到应用。
[Abstract]:Under the background of "Mass Entrepreneurship, Mass Innovation", the deployment and implementation of innovation-driven development strategy has become the focus of our government's work, among which, on the way to improve the ability of independent innovation in our country. Guiding and supporting innovation elements to bring together and promote the transformation of scientific and technological achievements into productive forces of industrial and academic cooperation plays an irreplaceable role. With the continuous promotion of industry-learning cooperation related work. Also found many of the problems: the depth of cooperation between industry and school needs to be deepened; The scientific and technological achievements of colleges and universities cannot effectively link up with the real market demand, thus affecting the quality of the cooperation between production and learning; Due to the lack of corresponding channels and information acquisition platform, it is easy to cause asymmetric information between the two sides of the school and enterprise. Generally speaking, as a demand-side enterprise, the mode of choosing cooperative university is single, mostly through the introduction of school fame or acquaintance. To a certain extent, colleges and universities will not be able to match the needs of enterprises, so that colleges and universities that are not well known but have strong professional ability in related fields lose the opportunity of cooperation. At the same time, enterprise resources can not be allocated reasonably. With the development of the Internet, the problem of information overload makes information screening and filtering become very important, how to use the advanced big data technology. Through the analysis of massive data, the problem of information asymmetry is solved for the two parties, and the appropriate partners are chosen to realize the more reasonable allocation of resources. To improve the quality and efficiency of industrial and academic cooperation has become the focus of this study. This paper analyzes the current situation of industrial and academic cooperation in colleges and universities through the analysis of the current situation and related literature research, combing the key indicators of the ability of industrial and academic cooperation in colleges and universities. From the perspective of enterprise requirements, the text preprocessing, feature extraction and keyword classification are carried out for the content itself. So that the enterprise needs can be more accurately divided into a certain field or category, and then combined with the characteristics of industry-learning cooperation, in the view of choosing partners, in the "basic scientific research capacity in colleges and universities". On the basis of the three key indexes of "the ability to transform the achievements of colleges and universities" and "the background of the cooperation between production and learning in colleges and universities", the paper continues to refine and enrich the indicators, and combines with the opinions of experts, adopts the 5-level scale assignment method. Through the least square regression fine-tuning, finally determined the weight of the indicators at all levels. Based on enterprise demand category information, the use of web crawler related technology. The introduction of asynchttpclient package and htmlparser package to obtain some of the colleges and universities to match the category of production and learning information, including papers and scientific research projects. This paper constructs a university recommendation algorithm model based on enterprise requirements, and then verifies the algorithm to a certain extent through an example. Finally, the object-oriented method is adopted. Based on UML, this paper analyzes and designs the university recommendation system oriented to the cooperation of production and learning, designs the interface of the system with the help of AxureRP prototype tool, and completes the display and interaction of Web pages. The recommendation model is further applied.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：G647;F279.2

【参考文献】