衡量信息过滤的效果主要有过滤速度、查全率和查准率三项指标。其中,过滤速

衡量信息过滤的效果主要有过滤速度、查全率和查准率三项指标。其中,过滤速度主要取决于信息过滤方法的算法复杂度。通过本文之前对几个常用过滤方法的分析可知,本文采用的基于k最近邻法的分类方法相对于贝叶斯法和支持向量机法的算法复杂度最低。而且,本文所设计的安全审计系统中用户需求文本集合很小,所以,k最近邻法中涉及的计算量就大大降低,从而保证了具有高速的过滤速度。

結果 (英文) 1:

Measure the effectiveness of information filtering filtering speed, precision and recall three main indicators. Among them, the filtration rate depends largely on the algorithmic complexity of information filtering methods. Before several methods of filtration is through this analysis, we use classification method based on k-nearest neighbor method in relation to Bayesian method and support vector machine method for minimum of computational complexity. Moreover, the user needs the security audit system in this article text is very small, so, k-nearest neighbor method will greatly reduce the amount of calculation involved in, thus ensuring a high filtering speed.

結果 (英文) 2:

Measure the effect of filtering the main filtration rate, recall and precision three indicators. Among them, the filtration rate depends on the complexity of the algorithm information filtering methods. Prior to this article through the analysis of several commonly used filtering methods shows that the method based on k-nearest neighbor method with respect to the Bayesian method and support vector machine algorithm complexity lowest used in this paper. Moreover, this paper designed security auditing system user needs text collection is small, so, k nearest neighbor method to calculate the amount involved will be greatly reduced, thus ensuring high-speed filtration rate.

結果 (英文) 3:

To measure the information filtering effect are mainly three index filtration speed, precision and recall. Among them, the filtration speed depends mainly on the information filtering algorithm complexity. Through this before on several commonly used filtering method analysis, this paper adopts the K classification method based on nearest neighbor algorithm is compared with the Bayesian method and the support vector machine algorithm complexity. Moreover, this security audit system design of user requirement text set is very small, so, k nearest neighbor computation involves the quantity method is greatly reduced, so as to ensure the filtering velocity with high speed.