基于作文自动评分理论的小学写作教学研究

发布时间：2019-03-13 12:35

【摘要】：作文自动评分简称AES (Automate Essay Scoring),指通过计算机软件对学生的作文(特指已输入计算机内的文本)进行评分。它解决了人工阅卷方式效率低、成本高、误差大、信息无法有效利用的问题,评分客观,且能够将学生写作信息反馈给教师。如今,它已经不是简单的作文评分工具,更广泛应用在写作教学领域。本研究中,笔者通过探寻支持作文能够实现自动评分的相关写作教学方面的理论,尤其是它的评分模型的建立过程,以获得对我国小学写作教学的启示。毋庸置疑,从古代到现代,我国小学写作教学研究领域发生了很多大变化,写作教学观念、理论、指导方式以及写作教学测评随时代发展不断更新。然而,我国小学写作教学仍存在着费时、费力且效率不高的现状,一个重要的原因就是小学写作评价的模糊性,造成教师对教学目标把握不明确,对于教什么没有清晰的框架,更没有合理的序列安排与指导过程。如今,国外作文自动评分系统的研究已日臻成熟,这项技术背后所反映的写作教学理论是值得我们借鉴与运用的。作文自动评分的基本原理是通过提高训练集(人工评分后的作文或者预先选定的范文)中人工评分的信度,然后建立评分模型,让机器能够模拟人工评分的过程。一些评分系统具有0.7以上的信度与效度,这证明它所建立的评分模型是一种科学的测量标准,为实现作文的有效评价提供了合理的依据。作文自动评分模型中的语言、内容、篇章结构评分三要素对我国写作能力结构分析、写作知识教学具有指导作用,有效促进小学写作教学过程中学生的基本写作能力与语言水平的发展。作文自动评分标准的建立不仅源于作文专家还有语言学家、心理学家等的支持,对于学生为什么写作有着更广泛的理解,对理解我国的写作教学目标有了更加清晰的方向。它的及时反馈性、提供的个性化学习环境,都揭示着写作教学将会走向新的变革。作文自动评分的研究是一个年轻的领域,是一个跨学科的领域。该理论虽然存在着不足,但对我国的小学写作教学目前的发展状况而言犹如一股新鲜的力量,有待我们进一步发掘。
[Abstract]:Automatic composition scoring (AES (Automate Essay Scoring),) refers to the grading of students' compositions (especially texts entered into the computer) by computer software. It solves the problems such as low efficiency, high cost, large error, unavailability of information, objective scoring and feedback of students' writing information to teachers. Today, it is no longer a simple composition scoring tool, more widely used in the field of writing teaching. In this study, the author seeks to support the composition can achieve automatic grading of the relevant theories of writing teaching, especially the establishment of its scoring model process, in order to obtain the enlightenment to our primary school writing teaching. There is no doubt that from ancient to modern, a lot of changes have taken place in the field of writing teaching in primary schools in China. The concept of writing teaching, the theory of writing teaching, the way of guidance and the evaluation of writing teaching have been updated with the development of the times. However, there is still a time-consuming, laborious and inefficient situation in primary school writing teaching in China. One of the important reasons is the fuzziness of primary school writing evaluation, which leads to the teachers' unclear grasp of teaching objectives and the lack of a clear framework for what to teach. There is no reasonable sequence arrangement and guidance process. Nowadays, the research on automatic composition scoring system in foreign countries has become more and more mature. The theory of writing teaching reflected in this technique is worthy of our reference and application. The basic principle of automatic composition scoring is to improve the reliability of manual scoring in the training set (composition after manual scoring or pre-selected model), and then establish a scoring model, so that the machine can simulate the process of manual scoring. Some scoring systems have more than 0.7 reliability and validity, which proves that the scoring model is a scientific measurement standard, which provides a reasonable basis for the effective evaluation of composition. The language, content and text structure of the automatic composition scoring model play a guiding role in the analysis of the structure of writing ability and the teaching of writing knowledge in China. It can effectively promote the development of students' basic writing ability and language level in the process of primary school writing teaching. The establishment of automatic scoring system not only comes from the support of writing experts, linguists, psychologists, etc., but also has a wider understanding of why students write and a clearer direction for understanding the teaching objectives of writing in China. Its timely feedback and personalized learning environment all reveal that writing teaching will move towards a new change. The study of automatic composition scoring is a young field and an interdisciplinary field. Although there are some shortcomings in this theory, it is a fresh force for the development of primary school writing teaching in our country, which needs to be further explored.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：G623.24

【相似文献】