基于数据分析的高考志愿决策模型研究

发布时间：2017-12-27 01:05

本文关键词：基于数据分析的高考志愿决策模型研究　出处：《山东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：每年的高考无疑是年度最受关注事件之一,牵动着亿万家庭的心。高考对于考生本人及家庭均有着重要的意义,学校、家庭、社会等对高考也极其重视,高考被很多考生及考生家长视为人生的一次转折。想要进入理想大学并非简单考出一个理想的成绩,志愿填报同样是决定能否进入理想院校或理想专业的关键因素,但志愿填报相对于高考而言其重视度不够,考生及其家长投入大量时间及金钱希望能在高考中取得好成绩,相对忽视志愿填报环节。此外,由于资源获取等方面的局限性,考生及家长不能获得足够的信息用于指导志愿填报。考生考出好成绩是前提,填好志愿是关键。因此志愿填报的地位对于考生来说应该仅次于高考,可以将其视为考生的“第二次高考”。十年寒窗,毁于一“报”,将成为考生终生的遗憾。在梯度志愿下,每年都有很多考生因为志愿填报的失误而未能走进理想院校。随着平行志愿的实行,平行志愿使考生拥有了更多的选择机会,但考生志愿决策失误的风险依然存在,同样会因为报考失误而无法走进理想的院校。有部分老师、考生及考生家长不能很好的对以往数据进行分析,造成填报志愿时出现偏差甚至是失误,导致高分考生没能进入自己心仪的大学甚至落榜。高考志愿填报是高考录取工作中不可缺少的重要环节,并被许多人称为“第二次高考”。然而,每年都有大量的考生由于无法有效的对志愿填报信息进行分析处理,只能简单的采用表面信息作为考生志愿填报的依据,导致高分考生最终无法进入自己理想的院校与专业,甚至落榜。文章针对报考中出现的问题,对数据进行了科学的分析与处理,帮助考生合理的选择院校与专业,减小高分考生进入不理想院校专业甚至落榜的可能性。由于不同年份高考信息的不一致性以及院校录取的波动性,导致考生无法根据现有数据正确的选择合适的专业与院校。本文提出了一种基于数据分析的志愿决策模型。首先通过分数—位次—分数的转换,将数据标准化;在此基础上,根据各院校不同年份招生人数的变动自适应的选择合适的数据分析系数,让历年分数信息更具有参考价值。然后根据院校波动类型,确定平滑系数,对院校录取分数的趋势变动进行二次指数平滑预测,确定模型中的关键点,建立录取概率模型;根据提出的模型计算院校录取概率,将院校依概率进行分层聚类,为考生确定报考院校的范围;最后通过考生对院校特征偏好的选取及评分,计算院校与考生理想院校之间的相似度,并以此为评价指标之一,进一步完成院校推荐。实验结果表明,该决策模型能够根据考生的分数、自身条件及选择偏好等因素有效的为考生提供志愿报考中的决策支持。
[Abstract]:Every year, the college entrance examination is undoubtedly one of the most important events of the year, affecting the hearts of hundreds of millions of families. The college entrance examination is very important for the examinees and their families. Schools, families and society also attach great importance to the college entrance examination. The college entrance examination is regarded as a turning point in life by many examinees and parents. Want to enter the ideal university is not simply to test an ideal result, voluntary reporting is also determined whether to enter the key factors of ideal professional institutions or ideal, but relative to the entrance voluntary reporting its importance is not enough, the candidates and their parents spend a lot of time and money, hope to get good grades in college entrance examination, the relative neglect of voluntary reporting links. In addition, due to the limitations of resource acquisition, candidates and parents can not obtain sufficient information to guide voluntary reporting. Examinees make good grades are the precondition, and the key is to fill in a good job. Therefore, the status of volunteering should be second only to the college entrance examination for examinees, which can be regarded as the "second college entrance examination" of the examinee. The long time, was destroyed in a "newspaper", the candidates will become a life-long regret. At the gradient of volunteering, many candidates have failed to enter the ideal college every year because of their voluntary mistakes. With the implementation of parallel volunteer, parallel voting makes candidates have more choice opportunities, but the risk of candidates' decision-making errors still exist, and it will fail to enter ideal colleges because of failing to register. Some of the teachers, students and parents can not be a good analysis of previous data, causing the deviation when completing the volunteer or even lead to mistakes, scores of candidates failed to enter their favorite college or even failed. The voluntary filling in the college entrance examination is an indispensable part of the entrance examination of the college entrance examination, and many people call it "the second national college entrance examination". However, there are a lot of candidates due to the lack of effective for voluntary reporting of information analysis and processing each year, simply by using the surface information as candidates voluntary reporting basis, resulting in high score candidates ultimately unable to enter their ideal colleges and professional, and even failed. According to the report the problems of data analysis and processing science, help colleges and professional choice candidates reasonable, reduce the scores of candidates into the ideal colleges even failed possibility. Due to the inconsistency of college entrance examination information in different years and the volatility of college admission, candidates can not choose the right majors and colleges according to the existing data. In this paper, a voluntary decision model based on data analysis is proposed. First, through the conversion fraction - ranking scores, data standardization; on this basis, according to the change of the college enrollment adaptive in different years the choice of the appropriate data analysis coefficient, let number information is more valuable. Then according to the fluctuation of college types, determine the smoothing coefficient, the change trend of admission two exponential smoothing prediction for colleges, determine the key points in the model, the establishment of admission probability model; calculation of college admission probability based on the proposed model, the probability of college level clustering, determine the scope for the candidates to apply for college; finally the candidates for the college selection and feature preference score, computing the similarity between colleges and universities and the ideal candidates, as one of the evaluation index, further complete the college recommended. The experimental results show that the decision-making model can effectively provide decision support for candidates in volunteer registration according to their scores, their own conditions and preferences.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：G637.3

【参考文献】