Research On Video Cloud Notes And Content Analysis In Internet Education

In today's society,Internet online teaching and classroom face-to-face teaching are the two main teaching methods in the education industry.Internet online teaching breaks the time and space constraints of face-to-face teaching in classrooms,providing learners with rich teaching resources and diverse learning platforms.However,while learning diversity changes,there is no breakthrough in learning aids,especially the video cloud notes.As a tool for learners and video resources to interact,video cloud notes provide convenient note-taking methods,perfect note-sharing mechanisms,and personalized note-taking services.A perfect video cloud note can play the potential of the content of the notes,tap the key points and difficulties of the teaching video,and help the learners to build a sound knowledge system.In the existing video cloud notes,learners use traditional keyboard input methods to record notes.At the same time,watching video and operating the keyboard will reduce learner learning efficiency.If you can integrate convenient and human-computer interaction speech recognition technology,it will be very Greatly improve this phenomenon.In the process of video learning,learners only pay attention to the memories and consolidate of the content of the notes,and the researchers do not know much about the analysis of the contents of the learner's notes,which is not conducive to the potential role of the content of the notes.Based on the above questions,the main contributions of this thesis are as follows.1.This thesis studies the DNN-HMM acoustic model in speech recognition technology,analyzes the composition and training of deep neural network model,combine hidden Markov model,establishes DNN-HMM acoustic model based on video cloud notes,analysis,designing and training them,and using the THCHS30 Chinese corpus and manually recorded experimental data to test the accuracy of the DNN-HMM acoustic model in video cloud notes speech recognition.The experimental results show that the word error rate and sentence error rate of the model are lower than the traditional GMM-HMM model,which can effectively improve the learning quality of video cloud notes.2.In-depth analysis of learner notes,using TF-IDF algorithm to extract note content hotspot.First of all,aiming at the shortcomings of traditional TF-IDF algorithm in the extraction of notes content hotspot,combined with the shortness and concise of notes,and the separate features of title and content,an improved algorithm based on notes content hotspot analysis TF-IDF-G is proposed,introducing note-word and position weighting factors in the algorithm.Then process the contents of the note,it includes Chinese word segmentation,part of speech tag and location acquisition.Finally,using the learning platform notes dataset to test and verify the correctness of improved TF-IDF-G algorithm in the extraction of notes hotspot.Finally,targeting the learning platform of the Network and Continuing Education Institute of Xidian University,implementing a video cloud note for online learning.Apply the trained DNN-HMM acoustic model to the voice dictation notes of the video cloud notes,the improved TF-IDF-G algorithm is used to analyze the contents of the notes and extract the hotspot of the notes.Experiment shows that the DNN-HMM acoustic model and the improved TF-IDF-G algorithm can effectively improve the video learning mechanism of Internet education,and improve learners' enthusiasm and efficiency.