Research And Implementation Of Automatic Document Review And Course Learning Big Data Platform

Education is the foundation of a country.In traditional education,teachers play a leading role,and students' learning is limited to specific places.Such a teaching method is not conducive to teachers' individualized teaching,but also hinders the development of personalized learning.In recent years,with the rise of new technologies such as artificial intelligence and big data,the education industry has also shown new vitality.Artificial intelligence makes the intelligent evaluation of education a reality.Through artificial intelligence,it can realize the automatic evaluation of homework documents and reduce the burden on teachers.At the same time,mining the hidden information in education big data can help teachers and students understand their own teaching and learning situation,and improve teaching quality.Therefore,this thesis studies and implements a course learning big data platform,which focuses on the establishment of an automatic document review model based on word movement distance and a course learning effect association analysis model based on big data.The automatic document review model studied in this thesis is mainly used for the evaluation of homework subjective questions.Its core is to calculate the text similarity using the artificial intelligence word level semantic analysis.An improved word mover's distance WMD algorithm – WMD-R algorithm is proposed.Compared with the original WMD algorithm,this algorithm uses the key word extraction algorithm Rake to calculate the word weight,and calculates the similarity of the whole text by calculating the word displacement distance between keywords.Rake algorithm is also improved.It no longer used only stopped words to divide words,but used Chinese word segmentation,stopped words and part of speech tagging together to divide words and get candidate keywords.Then through experiments,taking the average absolute percentage error and accuracy as indicators,WMD-R algorithm in this thesis is compared with the other three classical algorithms.Experimental results show that WMD-R algorithm outperforms the other three algorithms in terms of average absolute percentage error and accuracy.In addition to the above model research,this thesis also studies the implementation of a big data-based course learning effect correlation analysis model.Through the data mining algorithm Apriori,the correlation relationship between students' learning behaviors and learning outcomes is mined from a large amount of daily learning data stored in the course learns big data platform.The results showed that the students with good learning results performed well in various daily learning behaviors such as attendance,homework,experiment and class testing.Based on the research work of these two models,this thesis develops and implements a course learning big data platform,which solves the difficult problem of IT platform support in the new form of curriculum teaching,realizes a learning platform for learners,and the combination of online education and offline education provides support and services for teachers to carry out blended teaching and convenient course learning for students anytime and anywhere.