Analysis Of Students’ Learning Behavior Based On Cognitive Theory

With the continuous advancement of education information construction and the rapid development of online learning platforms,students’ learning styles are constantly changing.The rich learning resources of online learning platform create a self-learning environment for students,and also put forward requirements for students’ autonomous learning ability.In addition,engineering education is also paying more and more attention to the cultivation of students’ autonomous learning ability,which shows the importance of improving students’ autonomous learning ability for teachers.The cultivation of autonomous learning ability needs to pay attention to the development of students’ cognitive ability,but cognitive ability belongs to the internal thinking mechanism of students and is not easy to evaluate.Therefore,this study will make use of a large number of learning behaviors generated on the online learning platform of students to achieve explicit quantification of students’ implicit cognitive state from the perspective of metacognition and Bloom’s evaluation level,so as to provide new ideas for teachers to cultivate students’ autonomous learning ability.The main contents of this study are as follows:Based on the research status and research foundation at home and abroad,this thesis selects the students of the introductory course of the school of information and communication engineering of Beijing University of Posts and Telecommunications(hereinafter referred to as"introductory course")as the research object,and designs and implements the analysis scheme of metacognition and Bloom’s evaluation level based on students’ learning behavior respectively from the perspective of metacognition and the evaluation level of Bloom’s cognitive level.1.The analysis scheme of metacognition based on students’ online learning behavior mainly includes three aspects:(1)starting from the students’ autonomous learning process during the introductory course,16 online learning behaviors are constructed to form a set of learning behavior indicators;(2)metacognitive evaluation questionnaires including eight secondary dimensions of metacognition are compiled and distributed twice during the introductory course to obtain the change of students’metacognition;(3)for each metacognitive secondary dimension,independent sample t-test is used to screen out the online learning behaviors with statistical differences,and classic prediction model and self-developed INN neural network model are selected to predict the changes of students’each metacognitive secondary dimension.The experimental results show that the prediction accuracy of planning dimension and evaluation dimension of metacognition is 74.56%and 80.21%respectively,which proves that it is feasible to predict the development of metacognitive ability based on online learning behavior data.2.The analysis scheme of Bloom’s evaluation level based on students’learning behavior mainly includes four aspects:(1)Select the project mutual evaluation behavior in the introductory course to analyze the students’ evaluation level ability.A quality evaluation index system,WRPC,is constructed based on the review texts generated by students’mutual evaluation behaviors.The index system includes four indexes:Words,Relevancy,Professionalism and Credibility;(2)Based on the calculation method of WRPC evaluation index and multiple regression analysis,this thesis constructs the evaluation model EQ＿Model of mutual evaluation quality;(3)using text processing technology to segment the students’ comments,part of speech tagging and remove the stop words,and calculate the values of the four indicators of WRPC based on the processed text data;(4)The effect of EQ＿Model is verified and analyzed from two aspects.One is to evaluate the effect of the model by MAE and RMSE,the value of MAE is 0.6728 and the value of RMSE is 0.8563,indicating the effectiveness of the model.The second is WRPC value as the characteristic variables,select classic classifying text classification prediction model,and comparing with the predictions of EQ＿Model,according to the results WRPC evaluation index achieves good prediction results on Random Forests and EQ＿Model,and EQ＿Model’s effect is the best,show the rationality of WRPC evaluation index andn the validity of EQ＿Model.