A Survey Based On Rules And SVM For Education Information Classihcation

With the rapid development of network technology, various types of informationresources rapidly grow in the network, a large number of educational resourcesemerged into the network. The network education resources is increasingly becomingmore and more important way for students, educators and parents to access importantscientific workers an important source of information. However, the existing searchengines tend to search for information in a large number of irrelevant or uselesscontent, so how quickly and effectively obtain useful information and resources froma large number of information resources on education resources is the focus of thisstudy classification, while the automatic text classification technology is a network ofeducational resources automatic text classification key technologies.The main contents are as follows:1. Analyze the situation of the existing network educational resources and the mainnetwork behavior and needs, building classification system resources for basiceducation.(2) Based on the current existence of a large number of feature selection algorithm, inorder to properly determine the specific circumstances in which use of algorithmsneed to propose or judge can rely on the standard. This article reviews some of theliterature related to the basic feature selection algorithm, feature selection methodsand algorithms for the empirical comparison, and then propose a possible dependenceor determine criteria.3. Educational Resources affiliation exists between and parallel relationship, thispaper based on these relationships to be built as a hierarchy of web pages in HTMLformat discusses the main structural features (ie title, Anchor Text, meta) the impacton the web page classification, and proposed rule-based classification methods,experimental results show that the title and anchor text, etc. have a positive impact on the web page classification.4Build educational resources for the classification, this paper introduces the basictheoretical knowledge of SVM in the traditional SVM algorithm, based on the text forthe nonlinear separable problems outlier sensitivity of the results on the classification,an improved multi-class SVM algorithm (Weighted Multi-Class SVM), experimentalresults show that the algorithm is superior to multi-class SVM classification algorithmis better.5Classification Algorithm for Rule-based precision rate, the recall rate; improvedSVM algorithm precision rate, the recall rate problem, we propose a method ofcombining these two methods, experimental results show that the systemclassification effectiveness and efficiency are improved.