The Algorithm Analyses and Design about the Subjective Test Online Basing on the DOM Tree

The test system in this essay is a test system under the mode of B/S (Browser/Server). In the test system, the subjective item grading technology is always a problem that limits the computer scoring technology development. The subjective items generally require to answer the questions in a way of language description, Since different person has different way of thinking, different level of understanding and different way of describing, the answers cannot be unanimously the same. Here we´ll use the XML DOM object to do some research on the subjective item scoring problem. Solving the type of subjective items that has no unanimously same answers, such as short-answer questions, discussion questions etc. There are two factors that will affect the subjective item scoring: knowledge point and the nearness level. The unidirectional nearness algorithm in the fuzzy mathematics only focus on the keyword matching, but ignore the scoring of the knowledge point and the nearness level of the whole question´s answering. First we´ll analyses the DOM tree´s structure and the tree model, research the use of the navigation document tree, DOM tree object, attribute data´s reading and DOM tree DFS(depth first search) traversal method. And then discuss and build a automated scroing system´s workflow based on XML DOM tree. Design a scoring system algorithm that can solve the short-answer, discussion-question kind of subjective items problem.