| 摘 要: TFIDF算法作为一种加权算法，在信息检索和数据挖掘等自然语言处理领域发挥了巨大的作用。它的计 算模型相对简单，适合大数据并行计算，适用领域广泛，且拥有很好的解释性。基于以上这些特点，本文在TFIDF算法 基础之上，利用监督的学习，并通过引入加权因子和词贡献度，来修正TFIDF算法结果权值。利用这个算法可以在自然 语言处理中有效地提取特征标签，并且改进后的算法在这一细分领域具有极高准确度。 |
| 关键词: 自然语言处理；TFIDF；词加权算法；标签提取；监督学习 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 国家重点研发计划课题(课题编号:2017YFC080446). |
|
| Label Extraction Algorithm Based on Enhanced TFIDF |
|
WANG Jie,LI Xujian1,2
|
1.( 1.Shandong University of Science and Technology, Qingdao 266590, China;2.
2.The Key Laboratory of Digital Mine in Shandong, Qingdao 266590, China)
|
| Abstract: As a word weighting algorithm,TFIDF plays an important role in natural language processing such as information retrieval and data mining.TFIDF has relatively simple computational model,suitable for large data parallel computation,applied widely in many fields,and with good explanatory characteristics.Based on the above-mentioned characteristics,this paper proposes to amend the weighted results of TFIDF by means of supervised learning based on TFIDF algorithm as well as by introducing weighting factors and word contribution.This algorithm can effectively extract feature labels in natural language processing,and improve the degree of accuracy in this segmentation field. |
| Keywords: natural language processing;TFIDF;word weighting algorithm;label extraction;supervised learning |
