一种新的微博短文本倾向性检测模型

发布时间：2018-02-05 20:53

本文关键词： 微博短文本倾向性检测 Delaunay三角网聚类元胞遗传　出处：《情报科学》2017年11期 　论文类型：期刊论文

【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。
[Abstract]:[purpose / significance] with the increasing influence of Weibo's essay on social life. In this paper, a Weibo short text orientation detection model is proposed based on Delaunay triangulation clustering and cellular heredity. Firstly, the data collection and preprocessing of the detection method are described in this model. There are four stages of feature selection and index calculation, and a short text orientation detection model is established. Secondly, the Delaunay triangular network clustering and cellular inheritance are used to solve the above model. The key factors affecting the method are studied through simulation experiments. [results / conclusions] the results show that compared with the hierarchical structure method and neural network, the method is accurate. The correlation rate has good adaptability.
【作者单位】：义乌工商职业技术学院;
【基金】：浙江省自然科学基金(LQ13A040003)
【分类号】：G206;G254

【相似文献】