Research On On-Line Guidance Technology Of Autonomous Electronic Experiment Based On Artificial Intelligence Technology

Since the 21st century,the rapid development of modern science and technology,such as big data technology and information technology,has continuously promoted the reform and progress in the field of education.In particular,the application of artificial intelligence technology has brought new changes to traditional educational concepts,educational models and educational systems.In the upsurge of research on Intelligent education,experimental education has gradually moved to the road of intellectualization.On the other hand,during the epidemic of COVID-19 scourge,online and offline mixed experimental mode has become the mainstream experimental teaching mode in domestic colleges and universities.This experimental teaching mode solves many problems in traditional experimental teaching,but also brings many difficulties,such as the number of experimental users is much larger than the number of teachers,experimental guidance is difficult to cover all aspects,and so on.In order to improve the teaching quality of online experiments and improve the leading system of online experiments,this paper takes"Southeast Online Experiments"platform as the background,applies artificial intelligence technology to the guidance process of online experiments based on electronic circuit basic course experiments,and provides precise,standardized and personalized online experimental guidance for experimental users by using the user data and teaching resources of the platform.It provides technical support and research ideas for online experimental guidance system.In this paper,the identification method of user experimental circuit structure is studied first.User operation data obtained from"Southeast Online Experiment"background is used.Four artificial intelligence algorithms,Support Vector Machine,BP Neural Network,XGBoost and Cat Boost,are used to build online experimental category recognition models.The results of the models are analyzed and compared,and a more appropriate recognition scheme is obtained.The results show the accuracy,recall and₁-score of the Cat Boost algorithm in identifying user behavior is higher,the accuracy rate reaches 99.5%.Users can effectively identify the circuit structure of online experiment,which provides an effective target and basis for experimental teaching guidance.Secondly,this paper designs the guided interactive interface of the basic experimental process.After detailed study of the experiment from the perspective of experimental principle,instrument operation and circuit parameters,combining with the standard process of the experiment,the guided interactive process is designed to help users standardize operation in the process of the experiment and avoid some simple experimental failures.On the basis of this research,this paper also combines experimental textbooks and network knowledge base to sort out and summarize the knowledge points of electrical and electronic experiment faults,and uses the knowledge map construction technology to design the knowledge graph of electrical and electronic experiment faults.The guide interaction of basic experiment process and the Knowledge Graph of electrical and electronic experiment faults are important learning resources in experimental teaching guidance.After identifying the experimental circuit structure of the target user,standardized and personalized resource links are provided for them.Finally,this paper designs an experimental fault knowledge recommendation model based on Knowledge Graph and Recommended Algorithms,and the collaborative filtering recommendation algorithm is used to analyze the historical operation records of all users of the platform and calculate the similarity between the experimental operation errors.Then the Semantic Connotation of the experimental fault in Knowledge Graph is fused to introduce the experimental fault similarity based on knowledge map.The similarity obtained by the two algorithms is combined to create a suggestion list,which gives precise and personalized experiment direction for users.This model can better predict the user’s experimental behavior and greatly improve the accuracy of the recommendation by considering both user’s operation and connotation knowledge.