基于TextRank算法的项目标签智能化生成技术研究

Research on intelligent generation technology of project tag based on TextRank algorithm

为提高项目标签生成质量,研究基于TextRank算法的项目标签智能化生成技术。首先对项目文本进行预处理,包括括分词、去停用词、词性标注等三部分,然后通过计算各个节点之间的相似度,构建TextRank词图模型,最后通过迭代计算节点的权重(重要性得分),选出前K个关键词作为项目文本标签。结果表明:所研究技术应用下,提取出的关键词标签准确率、召回率、F值均要更高,由此说明所研究技术生成的标签质量更好,具有广泛的应用前景。

In order to improve the quality of item tag generation,the intelligent generation technology of item tag based on TextRank algorithm is studied.Firstly,the project text is preprocessed,including word segmentation,stop words removal and part-of-speech tagging.Then,the TextRank word graph model is constructed by calculating the similarity between each node.Finally,the first K keywords are selected as the project text labels by iteratively calculating the weights(importance scores)of nodes.The results show that the accuracy,recall rate and F value of the extracted keyword tags are higher under the application of the research technology,which shows that the tags generated by the research technology have better quality and broad application prospects.