Research And Application On Automatic Scoring Of Chinese Essays Based On HSK

Essay evaluation is a crucial component of language learning.The majority of essays are still currently examined manually.Although manual scoring is currently the most popular essay scoring technique,it has low efficiency and is readily influenced by the reviewer’s opinion.To increase the fairness and objectivity of essay scoring,many academics have started to focus on automated scoring technology.In order to reduce the influence of the subjectivity of the examiners on the essay scoring and improve the efficiency and accuracy of the essay scoring,this paper takes the essay of Hanyu Shuiping Kaoshi(HSK)as the research object,analyzes and studies the essay scoring model,aiming at improving the accuracy and efficiency of the essay scoring.The main work is as follows:(1)An essay scoring method based on the back-propagation network and the enhanced chimp algorithm is proposed.By incorporating the good point set,teaching,and memory methods,the original chimp optimization algorithm is enhanced.The back-propagation network’s model for evaluating essays is improved using the enhanced chimp method.The experiment’s results demonstrate that the model’s correlation between predicted essay scores and actual essay scores is superior to that of other scoring models.(2)An essay scoring method based on an improved African vulture algorithm for feature selection is proposed.The African vulture method is enhanced with the simplex and differential evolution strategies,which are then applied to data sets to reduce their number of features.Experimental results show that the algorithm can effectively reduce scoring features and improve the accuracy of essay classification.(3)An essay scoring method based on simultaneous optimization of SVM parameters and feature selection using an improved moth-flame algorithm is proposed.The feedback sharing mechanism and inertia weight factor are introduced in the position update stage of the moth-flame algorithm,and used to synchronously optimize the SVM parameters and scoring feature subset.The experimental results show that the model greatly reduces the number of scoring features and improves the accuracy of scoring classification.(4)Automatic essay scoring system with HSK-based is achieved.An automated essay scoring system based on HSK is created using the model from the study and analysis in the previous chapters.