一种新的微博短文本倾向性检测模型  

A Novel Detection Model for Tendency of Microblog Short Text

作　　者：牟式标[1] 陈志军[1] 

机构地区：[1]义乌工商职业技术学院,浙江义乌322000

出　　处：《情报科学》2017年第11期99-102,120,共5页Information Science

基　　金：浙江省自然科学基金(LQ13A040003)

摘　　要：【目的/意义】随着微博短文本对社会生活的影响日益增大,结合Delaunay三角网聚类和元胞遗传提出了一种微博短文本倾向性检测模型。【方法/过程】首先,该模型阐述了检测方法采取的数据收集、预处理、特征选择和指标计算四个阶段,并建立了短文本倾向性检测模型。其次,利用Delaunay三角网聚类和元胞遗传对上述模型进行求解。最后,通过仿真实验深入研究了影响该方法的关键因素。【结果/结论】结果表明,相比于层次结构法和神经网络,该方法在准确率、关联率等方面具有较好的适应性。【Purpose/significance】With the increasing of microblog short text’s impact on social life, a tendency detection method of microblog short text is proposed with delaunay triangulation clustering and Cellular Genetic.【Method/process】At first, the four stages of data collection, preprocessing, characteristics selection and index calculation are elaborated in the model, and the detection model of short text tendency is presented. Then, the model is solved to improve the prediction accuracy with delaunay triangulation clustering and Cellular Genetic. Finally, a simulation is conducted to research the influence factor.【Result/conclusion】The results show that, compared to hierarchical structure method and neural network, this method has good adaptability on accuracy rate, correlation rate and so on.

关 键 词：微博短文本 倾向性 检测 Delaunay三角网聚类 元胞遗传 

分 类 号：G254[文化科学—图书馆学]