【摘要】 特征词提取算法大多以选取单个关键词为主,存在词与词关联度缺失,文本主题表达不准确的问题,为此提出一种基于词共现的文本分类算法。通过计算词间共现率,利用TextRank算法构建共现图得到共现词组,建立文本向量表示模型,并利用SVM算法实现文本分类。实验证明,共现词作为文本特征项相对于传统单个特征词有更好的分类效果。更多还原

【Abstract】 Most feature word extraction algorithms are based on selecting a single keyword which exist some problems such as the lack of relevance between word and word,inaccurate expression of text themes.So this paper proposed a text classification algorithm based on word co-occurrence.By calculating the co-occurrence rate of words,the TextRank algorithm is used to construct the co-occurrence graph to get co-occurrence phrase.This design creates a text vector representation model and uses SVM algorithm to implement text classification.Experiments show that co-occurrence words as text feature items has a better classification effect than traditional single feature word.更多还原