Research On Visualization Of Measurement Data In Multimodal Program Debugging Process

In the field of programming education,behavioral measurement research in the process of programming debugging is of great significance for analyzing individual debugging behavioral characteristics,debugging strategies and eye movement patterns.However,relying only on traditional quantitative analysis methods,it is difficult to meet the analysis requirements of complex multimodal debugging behavior data.In recent years,more and more related researches have adopted visualization techniques to help analysts understand eye-tracking patterns in multimodal programming activities and gain insight into underlying cognitive processes.However,the existing related research is still in the exploratory stage in terms of thought process extraction,debugging intent recognition,and multi-person computing,and new visualization research schemes are urgently needed.This thesis takes the debugging and error correction task in the IDE as the measurement scene,and collects multimodal behavior data during the debugging process.A new sequence analysis method is proposed,which completes the pattern recognition of eye movement sequences,the fusion of debugging interaction actions and the aggregation of behavior sequences.Through the extraction and display of learning behavior trajectories,the observation of macroscopic thinking clues in the debugging process is realized,which effectively distinguishes the debugging thinking process of different individuals among groups,and makes up for the limitation of related research that it is difficult to extract group commonality and distinguish individual differences.Through spatial transformation,on the basis of code semantics,a visualization scheme is designed,and the performance of debugging behavior in abstract semantics such as tracing program control structures and function calls is observed,and multi-person superimposed operations and group common characteristics are realized under the same debugging intention.Using the above measurement techniques,this thesis carried out four batches of measurement experiments to measure the multimodal behavior of 58 first-year undergraduate students in 8 debugging and error correction tasks,and applied the above visualization scheme to the actual collected data set.The experimental results show that the visualization scheme in this thesis can be used to interpret thinking clues and debugging intentions,and compare the behavior trajectories during the debugging process.The research results of this thesis provide an effective graphic basis for teachers to observe students’ thinking activities,diagnose the difficulty of topics,and interpret students’ behaviors under specific topics.The research results of this thesis are of great significance for the extraction and analysis of learning behavior trajectories in the debugging process,and provide an effective graphic basis for teachers to observe students’ thinking activities,diagnose problem difficulty,and interpret students’ behavior under specific topics.