Analysis Of Examination Questions Stored In Word Based On Support Vector Machine

With the fast development of computer and communication technology, information society comes into being rapidly.Computer has become an indispensable important tool in people's work and life. Examination is a very important step in teaching process.Moreover,the making-up of test paper is me crux of the examination.If test paper-making still adopts traditional manual method, it is very hard for us to keep up with times.Using test paper-aided generation system for test paper-making has become an important research project in Computer Aided Instruction.This article mainly explores the method of the Word format of the paper's questions text block recognition. In the teaching process, a large number of papers are stored base on the Word form of file. As soon as the test system appears, we are facing a problem how to import the original papers in Word into item bank. this paper briefly describes the Word file storage structure, object model, and the program access. Start with the classic method of dealing with text characters automata theory papers by the papers selected for modeling, the model which easier to achieve in the program has the advantage of the principle of using the regular type, the disadvantage of it is that demand to mark recognition of text characters papers boundary and need to introduce boundary character. Boundary character is not present in the original paper, so we need to manually insert these borders which increase the workload of questions taken. Based on the theoretical basis of automata in the identification papers, considering the feature that the paper itself is the image, we make the text of the papers to the quantitative questions and give the special algorithm according to vector theory. We need not again to mark the border of text through the text block to identify the vector boundary questions which based on vector recognition papers. This method reducing manual input when the test questions into he amount of labor. In the Office software that comes with VBA environment, realizing the algorithm and acquiring experimental data. Finally, this article make a conclusion of the algorithm and prospect the follow-up.