综合素质评价是教育行业不可或缺的一个重要环节，也是中高等院校招生的一项重要指标。综合素质的评价结果由学生平时的点点滴滴所决定，言行举止、学习态度、为人处事、实践能力、身心健康等都是综合素质评价需要考察的方面。但是，老师、家庭对学生的表现可能存在理解上的偏差，导致综合素质评价得分具有很大的主观性，缺乏公平性。综合素质的评价方式也无法在社会上得到普遍的认同。随着信息技术的快速发展，依赖信息技术对学生平时的行为记录进行综合分析，得到公平公正的评价结果，减轻老师和家庭的压力，是非常有意义的。

本文设计与实现了一种基于信息技术的智能学生综合素质评价档案系统，以教育成长评价系统中学生各个学期平时的评价数据作为原始语料，通过语料清洗，利用感知机算法、词向量模型（Word2vec）、互信息和信息熵等自然语言处理（NLP）技术，达到将评价语句正确归类、计算综合素质素养相似度值、提取综合素质素养的关键短语信息的目标。

本文设计的评价档案系统开发使用了SpringBoot框架、Thymeleaf框架、JPA框架和MyBatis框架，数据库使用了MySQL。通过将评价语句正确归类，利用素养相似度值进行计算，以日期为单位作为横轴，绘制出学生每个素养的成长曲线。同时，通过旭日图的形式呈现出学生成长过程中的素养占比和关键信息。

本文的最后对系统的功能业务及使用的算法和模型做了验证和测试，结果表明，能够达到预期的性能要求和实现效果。同时，针对系统和算法模型的改进与优化提出了一些建议。

Comprehensive quality evaluation is an indispensable link in education industry, and it is also an important indicator for enrollment in secondary schools or colleges and universities. The comprehensive quality evaluation results are determined by the students' usual bits and pieces. Words and behaviors, learning attitudes, dealing with others, practical ability, and physical and mental health are all aspects that need to be examined in comprehensive quality evaluation. However, there may be deviations in understanding the performance of students by teachers and families, resulting in a subjective comprehensive evaluation score and lack of fairness. The comprehensive quality evaluation method also cannot be universally recognized in society. With the rapid development of information technology, it is very meaningful to rely on information technology to comprehensively analyze students' usual behavior records, to obtain fair evaluation results, and to reduce the pressure on teachers and families.

This paper designs and implements an intelligent evaluation archives system for student comprehensive quality based on information technology. The usual evaluation data of each semester of students in the education growth evaluation system is used as the original corpus. Through corpus cleaning, using natural language processing (NLP) technologies such as perceptron algorithm, word vector model (Word2vec), mutual information and information entropy, to correctly classify the evaluation sentence, to calculate a comprehensive quality literacy similarity value, to extract the key phrase information of comprehensive quality literacy.

The evaluation archive system designed in this paper uses the SpringBoot framework, Thymeleaf framework, JPA framework and MyBatis framework, and the database uses MySQL. By correctly classifying the evaluation sentences, using the literacy similarity value for calculation, and using the date as the horizontal axis, the growth curve of each literacy of the students is drawn. At the same time, the proportion of literacy and key information in the growth process of students is presented in the form of sunburst charts.

At the end of this article, the system's functional services and the algorithms and models used were verified and tested. The results show that the expected performance requirements and implementation results can be achieved. At the same time, some suggestions are made for the improvement and optimization of system and algorithm models.