Design And Implementation Of Teacher Portrait And Scoring System

The progress of information technology has laid the foundation for the arrival of the era of big data.Explosive information growth provides data basis for data mining.Various industries are actively applying data mining technology to explore potential information contained in existing data,and education is no exception.The application of data mining technology in education and teaching scenes go by the name of Educational Data Mining(EDM).EDM is an interdisciplinary subject involving computer science and technology,statistics and pedagogy.Teaching activity is a process of interaction between teachers and students.Teachers play a leading role in it,which is the key to improving teaching quality.Objective and fair evaluation of teachers' curriculum teaching is an important means to stimulate teachers' enthusiasm for work,guide teachers to improve teaching and improve teaching quality.In the Aspect of student feedback,although there are students' evaluation system in Colleges and universities,including scoring and writing evaluation,the processing of these data is still in the simple statistical analysis stage,and the average score is use to evaluate the teachers.This method implies some unfair treatment factors.Secondly,the common method of process text was read by teachers,which is very difficult for the academic affairs officer and teachers.The current teaching evaluation system does not show the teacher's portrait.It is difficult to understand its specific meaning only by the score returned.As a background,this paper makes data mining and Analysis on the text data of students' evaluation and teaching,and generates six dimensions of teaching design and content based on curriculum design and content,teaching methods,course management,course assessment,teaching attitude and learning harvest.In order to eliminate the hidden unreasonable factors,the new improved model is more fair,reasonable and convincing.In the process of processing the text,the author excavate the Aspect(Aspect)and Opinion(Aspect emotion words)of the text.The text processing method can be more detailed to understand the teacher's score in every Aspect,the precision rate(Precision)of the Aspect extraction part(Recall)and the recall rate(Recall)are above 80%.The accuracy rate of emotion classification in Opinion also reached 90%.The accuracy rate of the final teacher portrait is 70%.In the score improvement module,by observing the distribution of students' score,three modes of students' scoring are found: the overall distribution,the no area diversity,the outlier type.The authors use machine learning method to classify all students according to the scoring mode.The No.3 method has the best effect,accuracy is 99.18%.In order to improve the existing scoring model,this paper adopts two schemes,and takes different measures to deal with the three scoring modes.The two schemes have reduced the percentage of irrational factors to 0.08%,and the KL distance from the positive distribution is reduced from 0.24 to 0.05,and the correlation coefficient of the text is also raised to a significant correlation from low positive correlation.This study provides a more convincing method of teaching evaluation for teaching managers,and also produces a more accurate picture of teachers,presenting the feedback of students more effectively to the teaching managers and teachers to help them make the corresponding teaching adjustment,which is practical in this study.