The Research Of English Test Keywords Importance Evaluation Algorithm

Along with the increasing scale of the Internet, the network take people into the sea of information while providing convenience. Then the concept of keywords was born. Education network have a similar problem in providing English tests to customer. That is how to choose reasonable words as the basis of tests. This article researched and implemented a English test keywords classify system based on the idea of data mining. On one hand, the system could classify the keywords entered by users, on the other hand, the system will modify the evaluation model when the users changed the classify result if they are not satisfy with the result. First of all, the paper studies the current status and background of English test of high school, designs and develops a remote intelligent questions insert system based on MetaWeblog API. Then develops an English question keywords dynamic classification model using BayesNet algorithm. And at last an English question keywords importance evaluation system is designed and implemented.As implemented MetaWeblog API, question editors could use Microsoft Word to edit questions or paper and finish question insertion by clicking blog release button on the Word. Paper designs special regular expressions and uses them to segment and extract questions details. This way is different from regular as the editors could upload batch questions at one time. It provides a more efficient, accurate and convenient way.The paper chose three classification algorithms such as NaiveBayes, BayesNet and logistic. Three algorithms were compared on a larger number of experiments. At last BayesNet algorithm was selected as it has a relatively higher correct rate. The paper extracts three main key items of English test keywords under a lot of research on the English tests. And builds a relatively authority test data set by collecting the items values of keywords in a large English test database. On the basis of the data set, an English question keywords importance evaluation model is established. For the special characteristics of English test, this paper proposed a keywords dynamic weight redistribution a algorithms which based on the users’feedback statistics. The system allows user to modify the classification result and will record the modify logs. It will retrain the evaluation model when the user modify logs exceeds a certain threshold. So the evaluation model will be more accurate as it is a incremental learning model.