The Research And Application Of The Subjective Scoring Auxiliary

With the rapid development of computer technology, automated assessment system with computer have a wide range of applications, people use software to assess the objective questions include multiple-choice and fill-in, etc. But subjective questions such as the question answering is not used in applications. The main reason for this phenomenon is due to the limitations of theory of natural language understandingt. Generally the subjective questions is still marked by teachers manually, but the manual marking of the results will be affected by many subjective factors, so the results will be affected and impartiality. How to use the strengths of computing speed, high efficiency, high precision of computer into the application of the subjective questions has become an active research direction.The similarity computation we studied is focus on two levels:Chinese word and sentence, which are different but interrelated and going forward one by one. Firstly, a new method to compute Chinese word similarity based on Hownet is presented, this method is extensible and semantic-oriented; Secondly, we expanded the algorithm of edit distance algorithm, and regarded the chunk as its basic unit, and then Chinese sentence similarity can be computed easily; experiment results show that it is useful to calculate the similarity between sentence; finally, the automated assessment system is developed with methods presented above.When system running, the input are student answer and standard answer, with the process of segmentation and POS, extend the similar meaning of each word, and calculate the similarity using Hownet, which can be used to generate the similarity of sentence. Finally the score of student is given using the questing score and similarity.The study can be extended from the theory of natural language processing direction, the system has a wide range of practical significance.