Research And Realization Of Mathematics Automatic Marking System In Primary And Middle Schools Based On Machine Learning

In recent years,with the Big Data,Cloud Computing,Internet of Things and other new generation of information technology development.The way in which people acquire knowledge is also undergoing profound changes.Machine Learning,Wisdom Education,Education Big Data and other new ideas are changing the traditional Educational Ecology,Teaching Patterns and Learning Methods,but also to promote the Modernization of Education to the depth of information technology development.Examination is an important part of teaching behavior to understand the degree of knowledge and theoretical grasp of students and evaluate the teaching effect.The traditional artificial scoring has become the main burden of teachers,,which not only change the volume for a long time,but also prone to errors.Especially in the subjective items of the teachers' subjective factors in the scoring will have varying degrees of impact.Such as the degree of cleanliness of students' examination papers,whether the font is neat,teachers' were the mood and other factors.These may lead to the score to the points of the error,affecting the objective and impartiality of students' scores.Computer for automatic marking is not only fast and efficient,but also not affected from the subjective factors.Especially for large-scale and high-volume examination plays a more important significance.Therefore,research and development of computer automatic marking system.Not only can greatly reduce the burden of teachers' and help teachers focus on teaching innovation and scientific research,but also easy to true collection of students'learning data.Using Big Data Technology to dig the personality characteristics of students' and combine with Internet Technology and Cloud Computing Technology to promote the rapid development of Wisdom Education.This paper first reviews the development history of machine automatic marking at home and abroad,and analyzes the status quo of its technology development.For the difficulty of automatic marking.The idea of Deep Learning in Machine Learning is applied to the design of automatic marking system for subjective questions.And then build a hierarchical model of mathematical problems,and classify the topics.After the classification of the topics and then layered processing.In order to achieve a large number of students' examination information in depth mining analysis.To extract useful feature information to promote students' personalized and efficient learning.System for automatic marking for subjective questions on the standard answer input and fill in harsh questions.The system has been optimized for previous research.Just enter the topics'asked questions,titles' scores,subjects' standard answers and weights.Secondly,it is aimed at the problem of proper nouns in mathematics,but the case of the Forward Maximum Matching(FMM)algorithm is lower in the terminology of proper nouns.The Forward Maximum Matching Algorithm is improved and accuracy rate of word segmentation is enhanced.The theoretical knowledge of the close degree of Fuzzy Mathematics is applied to the similarity calculation of the keyword matching in this system.Finally,this paper uses the algorithm of binarization processing(Otsu Algorithm),corner detection(Harris Algorithm),straight line detection(Hough Algorithm)and graph parameter calculation.The parameter calculation of the geometric plots are realized.This paper presents a new model of mathematical automatic marking system for primary and secondary schools.And the development of primary and secondary mathematics automatic marking system for subjective questions.Through the actual primary and secondary mathematics tests.Which proves the practicability and accuracy of the system designed and realized in this paper.The system has a very good application to promote the value.