Research On Characterization Of Cognitive Features Based On Learning Behavior Data

On the way of realizing intelligent education,it is essential to make an in-depth analysis of how people learn.Cognition is the core of learning.For many years,cognitivist and constructivist scholars have tried to explain the conditions and nature of learning from the cognitive perspective.Cognitive features are important factors that determine the learning process and learning results of students.Fully understand the cognitive features of students,targeted teaching and resource push,etc.,will greatly stimulate and maintain learners' learning motivation and improve learning efficiency.Therefore,from the perspective of understanding the cognitive characteristics of students,devoting to open the valve of student learning is one of the key factors to improve the current online learning support.Although behaviorist representations have died due to factors such as the immaturity of data collection and analysis techniques,the dogma and pursuit of behaviorists(that the appropriate dependent variable in psychological research should be observable and verifiable behavior)have always been the expectation of research related to cognitive features.Especially in recent years,with the development of learning analysis technology and the demand for intelligent education,it is increasingly urgent to explore the representation of cognitive features from the perspective of behavior.In the context of the increasingly mature demand for behavior-based research paradigm and behavior-based research technology support,this research investigated three questions: "What are cognitive features based on behavioral perspective? How to characterize cognitive features based on students' learning behavior? What are the cognitive features based on student learning behavior data?" It is hoped that from the perspective of behavior,the theory and method are enriched and perfected for the representation of cognitive features.This research takes two courses of college students in a university in Shanghai as the research object,and tracks and collects the data of students' learning behavior through the online learning platform Moodle.This paper deconstructs students' learning cognitive process from three aspects: theory driven,data driven and expert method,and then extracts,excavates and verifies the framework of cognitive features characterization.On this basis,in order to further explain the application context,data collection design,and cognitive features representation method of the cognitive features framework,this study conducted an explanatory study on the framework based on students' behavior data.Specifically,the main research conclusions of this study are as follows:First,this study interprets the essence,situation,characteristic,representation and value of cognitive features from the perspective of deconstruction.The essence of cognitive features is the difference of cognitive styles,the situation of its expression is the autonomous learning situation,its characteristic is the process,hierarchical presentation and stability,and its representation has the characteristics of stage and behavior.The value of clarifying students' cognitive features lies in assisting teachers and administrators to implement "learning aid according to ability".Second,the study develops and verifies the representation framework of cognitive features from three perspectives: theory-driven,data-driven,and expert methods,and concludes that the cognitive features representation behavior framework includes:(1)Students' perceptual learning behavior(selection of learning materials),information processing(timing of learning,information preservation,information categorization/integration),and autonomous problem solving(testing behavior)in the learning activities of independent knowledge construction;(2)The exchange learning behavior between students and teachers and between students in exchange learning activities.(3)Collaborative learning behavior of students in groups in assigned task learning activities,including participation in peer discussion,feedback,and evaluation?Thirdly,the representational interpretation of cognitive features of students' perception,information processing and problem solving was studied,and it was found that:(1)Different perceptual and cognitive features make students have different preferences and preferences when choosing the presentation mode of learning materials;(2)There is a significant difference between holistic and analytical learners in the time sequence of learning,which is consistent with the existing research conclusion,that is,compared with analytic and holistic learners,most of their learning order is not step-by step,but according to their own pace.The study also verified the stability of cognitive features,and found that the behavior sequence of students had periodicity,that is,the trend and starting point of the behavior sequence had similar characteristics in a period of time;(3)In the exchange task,there were statistically significant differences in the number of students looking at the notice,the number of teachers asking for help and the number of forum posts asking for help.Type 1students are more independent when encountering problems;in addition,this type of students have a lower willingness to interact with teachers and other students.Type 2is the opposite.They are more dependent on other members of the learning community and the learning environment,and they are more willing to ask others for help when they encounter problems.Moreover,the students of Type 1 performed better in the process of self-directed problem solving than those of Type 2,with longer answering time,fewer answering times and higher accuracy.Finally,the paper explained how to effectively promote the generation and collection of cognitive features data,and given suggestions from the aspects of tool design,data collection and data analysis.In summary,the innovation and value of this study are as follows:(1)This study starts from the behavior data which can more truly represent the current situation of students,constructs and analyzes the cognitive features of learners,and opens up a new paradigm of cognitive features research to some extent.(2)The framework of characterizing cognitive features from activity-behavior-operation is developed.The development and interpretation of the framework provide theoretical support for characterizing cognitive features.(3)Multi-perspective and multi-mode data analysis methods,such as statistical analysis(difference),sequence analysis,machine learning(clustering algorithm),social network and other ways to collect data,detailed research processes and methods to provide guidance for relevant research data collection and analysis methods;(4)The results of cognitive features obtained from the research have certain guiding significance for teachers' teaching.Meanwhile,the research results have practical guiding significance for the construction of intelligent learning system in the future,for the classification based on students' cognitive features.