【作者】 米晓芳； 秦洋； 王立宏； 宋宜斌；

【Author】 Mi Xiaofang Qin Yang Wang Lihong Song Yibin (Institute of Computer Science and Technology,Yantai,264005)

【机构】 烟台大学计算机学院；

【摘要】 通过分析潜在语义索引的各种模型提出一种新的潜在语义差异模型。将医学网页中的文本抽取出来并分别采用全局模型、局部模型和差异模型进行表示,利用 SVM 算法进行分类并计算分类正确率和 F₁指标。实验发现:在差异模型表示下,分类正确率和 F₁指标较全局模型有明显提高,某些情况下较局部模型分类结果也有提高。更多还原

【Abstract】 A new latent semantic difference model is proposed on the base of various LSI models.The text in medical webpage is extracted and represented by global LSI,local LSI and difference LSI respectively.SVM algorithm is employed to classify the feature vectors of testing collection and the categorical accuracy and macro-average F₁ are calculated.Simulating experiment illustrates that difference LSI is outstanding in increasing accuracy and macro-average F₁ when compared with global LSI,and obtains some improvement on accuracy and macro-average F₁ when compared with local LSI.更多还原