模块化网络主题语义分类算法的设计与仿真

Design and Simulation of Modular WebTopic Semantic Classification Algorithm

由于目前已有算法没有对网络主题语义流行度进行计算,构建流行度序列,导致分类结果不理想,运行时间增加。提出一种基于特征序列的模块化网络主题语义分类算法,通过Ochiia系数方法将高频关键词同时出现的次数转换为网络主题语义相似性。将语义相似性较强的网络主题聚集到一个簇中,形成主题簇。根据特征序列获取网络主题簇的平均热度和描述词,以此为依据构建流行度序列。采用特征序列对全部序列进行分类,实现模块化网络主题语义分类。仿真结果表明,所提算法能够快速准确完成模块化网络主题语义分类。

Because the existing algorithms do not calculate the semantic popularity of network topics and construct the popularity sequence,the classification results are not ideal and the running time is increased.Therefore,a modular classification algorithm for network topic semantics based on feature sequence was proposed.The Ochiia coefficient method was used to convert the number of simultaneous occurrences of high-frequency keywords into semantic similarity.The topics with strong semantic similarity were gathered into a cluster to form a topic cluster.According to the feature sequence,the average clout and description words of network topic clusters were obtained.On this basis,the popularity sequence was constructed.Finally,the feature sequence was adopted to classify all sequences.Thus,the modular classification of network topic semantics was achieved.Simulation results show that the proposed algorithm can quickly and accurately complete the semantic classification of modular network topics.