节点文献
基于多源信息融合的本科生综合素质评价研究
Research on Evaluation of Undergraduates’ Comprehensive Quality Based on Multi-source Information Fusion
【作者】 方宝红;
【导师】 杨保安;
【作者基本信息】 东华大学 , 管理科学与工程, 2011, 博士
【摘要】 随着高等教育进入大众化,全球化,价值多元化时代,如何评价专业不同、禀赋各异的本科生的素质,是一个复杂的问题。仅凭个人的精力、知识、经验和智慧已无法掌握所有必要的与本科生素质相关的知识和信息来做出合理的评价。评价者在评价时需要综合多种来源的信息,由多位专家以及与本科生培养过程中密切相关的人员组成评价群体进行评价,依靠群体的智慧、信息和知识来解决问题。由于受到评价者知识结构、评判水平和个人偏好等主观因素及本科生综合素质本身的模糊性、不确定性和复杂性的影响,评价信息具有模糊性和不确定性。本科生综合素质评价面临的主要问题包括:第一,评价指标的确定及评价主体单一的问题。不管评价的对象是一年级新生还是高年级本科生或者毕业生,不管评价主体是教师、辅导员、管理人员还是校外人员,评价指标没有区分,指标权重没有区分,因而评价结果不符合本科生综合素质的个性化评价要求;第二,评价过程是静态的,不能考虑本科生的成长过程,对本科生的多次评价没有反馈,因而无法体现本科生的成长历程,更无法提出针对性的激励措施。第三,评价信息来源单一,评价结果无法真实反映本科生的综合素质水平。针对以上问题,本文在考察了现有本科生综合素质评价系统的基础上,提出了本科生综合素质评价框架结构,进一步运用信息融合理论、Agent技术等研究综合评价过程中的多源数据获取、数据融合、评价指标及权重、个性化综合评价模型,形成了本科生综合素质评价的完整理论体系,为高校进行本科生综合素质管理提供理论支撑,所得到的重要结论对引导高校个性化素质教育、提高高校素质评价与管理水平具有较强的实践指导意义。论文的研究内容如下:一、通过对本科生综合素质评价目标、评价方法、评价指标的文献综述,结合高校本科生综合素质评价管理的现状,分析现有评价体系中存在的问题,提出评价人才的标准有多种类型、多种规格、多种层次,避免千人一面,充分发掘本科生潜能,主张多元评价主体和多元价值评价理念。二、提出建立个性化、动态的综合素质评价体系,明确针对不同评价主体、不同评价对象进行个性化定制评价的必要性和可行性。进一步从模型、应用和实施三个维度建立了本科生综合素质评价的框架结构。第一个维度是模型维度,主要说明论文研究的思路,包括多源信息的分析;多源信息的信息融合方法建立相应的数据库、模型库;通过二元语义融合模型确定评价指标、指标的权重,并进一步建立个性化可定制的评价模型。第二个维度是实施维度,采用多Agent技术,建立包括信息Agent、信息融合Agent、评价Agent、管理Agent和界面Agent的多Agent系统,对个性化评价过程进行模拟。第三个维度是应用维度,主要考虑应用的主体、客体(对象),对不同的主客体可以进行不同评价模式的定制。三、本科生综合素质评价的数据来源是多源异构的,既有来自于各类不同人员的知识,又有来自于不同信息系统的数据信息;评价方法是多样的。因而采用多源信息融合方法将多源数据、多种方法进行综合建立信息融合数据库、模型库等,为评价提供信息基础和技术支持。四、在本科生综合素质评价的研究中,出于本科生综合素质评价过程中有部分信息具有主观性和模糊性的特点,为了避免信息损失,首先将客观评价、主观评价等评价信息进行标准化处理,为进一步的数据分析提供有效的目标数据;然后借助二元语义方法、信息融合算子进行群体评价信息的集结;最后计算得出了综合评价的指标权重。五、建立了可定制的本科生综合素质评价模型,并详细阐述了个性化评价指标的选取及个性化评价过程。对本科生综合素质进行动态的和个性化的评价,避免了千人一面、千篇一律的评价,体现了以本科生为中心的个性化评价的思想。本文的创新性体现在以下三个方面:一、重视本科生个性、以人为本的理念,将多源信息融合理论和多Agent技术引入到本科生综合素质评价研究中,建立了模型维度、应用维度和实施维度的三维立体的本科生综合素质评价框架结构,为本科生综合素质评价体系的建立提供了思路。二、采用信息融合方法将多源数据、多种评价方法进行综合,并采用多元信息融合算子和二元语义方法给出了综合素质评价的专家群体评价意见,并确定了综合评价的指标权重,为综合评价提供了理论依据。三、考虑不同的评价主体和评价对象,建立了个性化的可定制评价模型,针对不同类别的本科生培养的不同阶段和不同的评价目标,用不同的评价模型、评价准则来进行评价。该评价模型不仅可以对本科生的全面素质培养起到导航作用,也使高校的素质教育和管理工作更具针对性,主动性和实效性。
【Abstract】 It is a complicated problem to evaluate the quality of the undergraduates according to their different majors and characteristics when higher education is becoming popular, globalized and value-diversified. It is unable to master all the necessary and relevant knowledge or information to evaluate the quality, and make a reasonable evaluation depending on personal energy, knowledge, experience and wisdom. Evaluators have to integrate multi-source information in the procedure of evaluating, from experts as well as undergraduates training process is closely related to the composition of the evaluation group to evaluate the wisdom of relying on groups, information and knowledge to solve problems. Evaluators have to deal with information from different sources. It was used to evaluate knowledge of the structure due to the evaluators, judging the level and personal preferences and other subjective factors and the overall quality of undergraduates by the evaluation group which made of experts and relevant personnel closely to undergraduates training.The Main problems on evaluating the comprehensive quality of the undergraduates were divided into three aspects. Firstly, the problem of establishing stable evaluation index system and the singleness of evaluation objects. Whether the object of evaluation is a freshman, upperclassman or graduates, regardless of evaluation of the main teachers, counselors, administrators or school personnel, there is no distinction for index indicators and therefore not suited to evaluate the result of quality to personalized evaluation requirements. Secondly, the evaluation process is static and their growth process was not considered, and no feedback is given for the result of the evaluation. So the result could not reflect the growth process of the undergraduates and it’s hard to put forward the pertinent incentive policies. Thirdly, the evaluation result is too simple to reflect of the true level of quality of students.Given the above problems, after examined the existing undergraduates" comprehensive evaluation system, the paper proposed the framework for Comprehensive Evaluation of Undergraduates, furthermore. Information Fusion Theory. Agent Technology were also used in the evaluation process of comprehensive multi-source data acquisition, data fusion, assessment index and weight, personalized comprehensive evaluation model were made of the complete theoretical system for comprehensive evaluation for undergraduates’quality, which provided theoretical support. Important conclusions were used to guide the university personal qualities education and improve the quality of evaluation and management with strong practical significance. The main contents of the paper are as follows:Firstly, existed evaluation system problems were analyzed on the quality of undergraduate’s through the evaluation objectives, evaluation methods and the literature review of evaluation index, combined with the status of evaluation management of the quality of university undergraduates, it was pointed out that evaluation criteria for a variety of types of talent, a variety of specifications, a variety of levels, was to avoid using the same one for thousands of students, to fully exploit the potential quality of students, and appraised subject pluralism, appraised in contents and appraise method variety.Secondly, the paper proposed the establishment of personalized, dynamic and comprehensive quality assessment system, a clear evaluation of different subject, different evaluation objects to customize the necessity and feasibility evaluation. Furthermore a framework was built for comprehensive evaluation from the model, application and implementation of the three dimensions of undergraduates. The first dimension is the model dimension, the main idea of that thesis, including the analysis of multi-source data source; multi-source data fusion method to establish the corresponding database, model base; determined by binary semantic fusion model assessment index, index weight, and further establish the evaluation model customization. The second dimension is the implementation of the dimension. multi-Agent technology, the establishment of multi-Agent system including Information Agent, Information Fusion Agent. Evaluation Agent, Management Agent and Multi-Agent System Interface Agent, individualized evaluation process simulation. The third dimension is the application of the dimension, mainly considering the application of the subject and object. The main object of different evaluation models can be customized for different objects.Thirdly, considered the quality of the comprehensive evaluation of the data source is multi-source heterogeneous, both from the knowledge of all kinds of different people, but also from different information systems data; evaluation method is diverse. Thus a multi-source information fusion method was used on multi-source data, a variety of methods to establish a comprehensive database of information fusion, model base, provided information infrastructure for the evaluation.Fourth, in the study on the quality of evaluation of the undergraduates, considered a lot of subjective quality assessment information, fuzzy information, it were avoided to loss of information, first objective evaluation, subjective evaluation of standardized evaluation of information for further data analysis were a valid target data; and then 2-tuple linguistic representation, information fusion operator were used for the evaluation of the quality of the views of experts on group decision making; the paper calculated a comprehensive evaluation of the index weight at last.Fifthly, a customized evaluation model on the quality of students was made, so the detailed evaluation of the individual and the individual evaluation of the selection process were described. The unbalanced, unified assessment was avoided and the student-centered evaluation based on personalized ideas was reflected for using dynamic quality of undergraduate and personalized evaluation.The innovations were reflected in the following three aspects:Firstly. the emphasis on student’s personality, people-oriented concept, the introduction of the theory of Multi-source Information Fusion and Multi-agent Technology into the Comprehensive Evaluation of the quality of undergraduates, the establishment of the model dimension, the application of three-dimensional dimensions and implementation of the dimension framework for Comprehensive Evaluation of undergraduates, which provided a guideline to establish comprehensive quality evaluation system.Secondly, a comprehensive evaluation of the expert group decision quality advice and a comprehensive evaluation system to determine the index weight was given to the usage of information fusion methods to multi-source data while a comprehensive variety of evaluation methods and usage of multiple information fusion operator and 2-tuple linguistic representation method, which provided a theoretical basis for evaluation.Thirdly, considered the different evaluation of the main objects, the establishment of a personalized evaluation model tailored was used for different types of education of undergraduates at different stages and different evaluation objectives, with different evaluation models, evaluation criteria to be evaluated. The evaluation model was not only used for the quality of undergraduates navigation training, but also for the quality of university education and more focused management, which would be more proactive and effective.
【Key words】 Quality; Group Evaluation; Personalized Evaluation; Information Fusion; 2-tuple Linguistic Representation; Customized;