| Educational reforms pay more and more attention to the individualized cultivation of students.In order to promote the comprehensive and individual development of students,teachers need to formulate individualized teaching strategies and evaluation methods according to the different learning characteristics of students.As an important factor influencing students’ academic performance,teacher comments can provide guidance for students’ learning on the one hand,and stimulate students’ enthusiasm for learning on the other hand.The research on teacher comments is becoming more and more mature and it starts to pay attention to the emotional development of students,but it has not been combined with the individualized cultivation of students.This research adopts the sentiment analysis method to extract the emotional factors related to students’ learning power and teacher comments types,aiming to explore the best comment type output for students with different learning status and learning needs.First of all,this research has determined the classification perspective of teacher comments(focus,valence,presence or absence of suggested words)and the specific indicators of students’ learning power(learning motivation,learning ability,and learning perseverance)through literature review;then,sentiment analysis is used to extract the teacher’s evaluation information of the student’s learning power development level and the teacher’s comment type information from the teacher comments database and then the teacher comment sentiment dictionary is constructed;then the scoring formula is designed based on the weight value and polarity value of the sentiment words to calculte students’ learning power acore and determine the type of teacher comments.Finally,Two-Way ANOVA is applied to analyze the impact of different types of teacher comments on the achievement performance of students with different learning power levels.There are mainly three research results.Firstly,the sentiment analysis method of teacher comments is explored and practiced,and the practical results have been proved to have high accuracy;secondly,student factors are incorporated into the teacher comment research system and we explore themost suitable types of comments for students with different learning power levels,which can provide guidance for the feedback behavior of primary and middle school teachers;thirdly,based on the sentiment analysis process,this research has developed a small tool to assist teachers in checking their own comment types and students’ learning power level by inputting comments with one click. |