Online Learning Behavior And Learning Effectiveness Data Visualization Research And Application

At all levels of the education system,student learning is no longer confined to the classroom desk;the Internet has brought unprecedented and novel ways of teaching and learning,allowing students to enjoy access to quality education at any time and place.However,against the backdrop of a plateau in epidemic prevention and control,online network education,as the focus of reform in the modern education system,needs to be explored in depth as to whether it is effective in student achievement growth.The primary data source for this paper is the behavioral data from students’ interactions with online learning platforms.This paper focuses on the registered students of the online education platform of the College of Continuing Education of Southwest University of Science and Technology as the research object to study student learning behavior data.The K-Means clustering algorithm summarizes and aggregates online learning behavior states to clarify each student’s current learning state and whether it is similar to most students’ learning states.The paper considers the large variety of learning behavior features and the unclear learning state classification categories for applying the K-Means algorithm in learning state classification.The paper uses the relatively farthest strategy to select the initial clustering centre for the problem that the K-Means algorithm is sensitive to the initial clustering centre.At the same time,the multi-dimensional learning behavior features have different degrees of influence on performance.K-means treats the weights of learning behavior features that impact performance regardless of their size as the same.The accuracy of learning state clustering is significantly reduced,so the weights of learning behavior feature with small influence are set low in the improved algorithm.The experimental analysis shows that the enhanced algorithm has significantly improved computing efficiency,computing performance and computing accuracy so that the management can more accurately control the current learning status of students.Bp neural network is combined with students’ learning states to predict learning performance.The online learning management system is used to visualize the learning status and predict students’ learning performance.Student performance is a direct response to the learning status.After clustering the learning status,the prediction of student performance can reflect the current learning status of students more intuitively.The processed online learning behavior data is used as the input of the BP neural network,and the grades are used as the prediction output.At the same time,Genetic Algorithm(GA)is introduced to solve the problem of the BP neural network quickly falling into local miniaturization.The optimal initial weights and thresholds are used as the initial weights and thresholds of the BP neural network through the optimization-seeking ability of the genetic algorithm to construct the GA-BP optimization model.Adaboost is used to fuse multiple GA-BP neural networks as weak classifiers and output a robust classifier prediction model with a weighted summation of multiple weak classifiers.The final experimental results show that the improved network performance of the integrated algorithm is better.They are improving a more accurate algorithmic model for predicting student performance.Using an online learning behavior state classification clustering model and combining it with data visualization technology,we build an online learning behavior monitoring system and visualize the online learning behavior monitoring results according to the learning behavior analysis module.