Automatic subjective-question scoring system and method based on semantic similarity interval

Abstract

The invention discloses an automatic subjective-question scoring system and method based on a semantic similarity interval. The automatic subjective-question scoring method based on the semantic similarity interval comprises the following steps that the score S<total> of a question to be scored is initialized; the block length L is defined; a reference answer is segmented into a plurality of blocks with the lengths being not larger than L, and a reference answer block set R is formed; an answer to be scored is segmented into a plurality of blocks with the lengths being not larger than L, and a block set T of the answer to be scored is formed; the set R and the set T are compared, and the semantic similarity S<RT> of the set R and the set T is worked out; the S<RT> is mapped into the similarity interval, a score is recorded as S<final>, and scoring is finished. The automatic subjective-question scoring system based on the semantic similarity interval comprises an answer terminal, a subjective-question paper inspection module and an examination result generating module. According to the automatic subjective-question scoring system and method based on the semantic similarity interval, as for automatic paper inspection and scoring links of subjective questions, grading can be conducted on answers to the subjective questions according to semantic similarity threshold values, and therefore final scores can be given through score restriction and similarity scoring of all grades.

Description

technical field [0001] The invention belongs to the technical field of education and teaching examinations, and in particular relates to an automatic scoring system and method for subjective questions based on semantic similarity intervals. Background technique [0002] At present, in the existing examination forms, subjective questions (such as essay questions, etc.) account for a certain proportion of the examination papers. The evaluation and judgment of the answers to subjective questions will directly affect the final score of the examination paper. Therefore, it is required to have more reasonable normative constraints on the scoring process of subjective questions. Manual marking and scoring methods are susceptible to interference from multiple factors, which makes the scoring results somewhat subjective. Based on intelligent information processing methods, especially Chinese information processing technology, research on objective scoring methods and models for the ...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the embodiments of the present invention is to provide an automatic grading system and method for subjective questions based on semantic similarity intervals, aiming at solving the problem that manual marking of subjective questions in existing examinations is mainly subjective, and it is difficult to ensure fairness. The problem

Images