基于改进模糊语法增量式算法的文本分类方法

发布时间：2018-05-04 14:16

本文选题：文本分类 + 机器学习　；参考：《计算机应用研究》2017年11期

【摘要】：针对现有很多文本分类算法必须进行训练—测试—再训练的缺点以及通用模型的语法表现度较差等问题,提出一种改进的模糊语法算法(IFGA)。根据一些选取的文本片段建立学习模型;为了适应轻微变化,采用增量式模型,将选取的文本片段转换到底层架构中,形成模糊语法;利用模糊联合操作将单个文本片段语法进行整合,并将所学习的文本片段转换成更加一般的表示形式。与决策表算法、改进的朴素贝叶斯算法等进行了两组对比实验,第一个实验结果表明,IFGA和其他机器学习算法性能并无明显差异;第二个实验结果说明,增量式学习算法比标准机器学习算法更加具有优势,其性能较平稳,数据的尺寸影响更小。提出的算法具有较低的模型重新训练时间。
[Abstract]:In view of the shortcomings of many existing text classification algorithms which must be trained, tested and retrained, as well as the poor grammatical representation of the general model, an improved fuzzy grammar algorithm (IFGAA) is proposed. The learning model is established according to some selected text fragments, and in order to adapt to slight changes, the incremental model is adopted to convert the selected text fragments to the underlying structure to form fuzzy syntax. A fuzzy joint operation is used to integrate the grammar of a single text fragment, and the learned text fragment is converted into a more general representation. Compared with the decision table algorithm and the improved naive Bayes algorithm, the first experiment shows that there is no significant difference between the performance of IFGA and other machine learning algorithms, and the second experiment shows that, The incremental learning algorithm has more advantages than the standard machine learning algorithm, its performance is stable, and the size of the data has less influence. The proposed algorithm has lower model re-training time.
【作者单位】：湖南环境生物职业技术学院信息技术系;南华大学计算机学院;
【基金】：国家自然科学基金资助项目(61300234) 湖南省教育科技计划资助项目(13C243,12C1056)
【分类号】：TP391.1

【相似文献】