基于改进K—Means算法的教学反思文本聚类研究

对教学反思内容的准确评估是教师基于教学反思过程提升其专业能力的重要保障。基于改进的K-Means算法对相同主题的教学反思文本进行聚类，通过给定初始聚类中心K的取值范围使其可以在给定范围内自动增加，在聚类过程中加入相似度阈值以限定文本间相似度的取值范围，实现对教学反思文本的分类和对自我反思文本的定位。实验结果表明改进的K—Means算法在反思文本聚类的准确率和稳定性方面比传统算法有所提高，且能根据教学反思内容准确地进行自动分类。

Research on Text Clustering of Teaching Reflection Based on Improved K-Means Algorithm

An accurate assessment of teaching reflection content is an important guarantee based on teachers＇ teaching reflection process to enhance their professional capabilities. Clustering the same theme of the teaching reflection text based on an improved K-Means algo- rithm, through given the initial cluster center K a value ranges, so that it can be automatically increased within the given range, during the clustering process, similarity threshold is introduced to limit the reflection texts＇ similarity ranges, realizing the teaching reflection text classification and the self-reflection text classification. The experiment result indicates the improved algorithm has a higher accuracy,better stability, and can accurately automatically classify according to the teaching reflection content.