Research And Application Of Keyword Extraction On MOOC Video

With the rapid development of Internet technology and mobile communication technology,Massive Open Online Course(MOOC)is rising rapidly all over the world.Mass instructional videos are provided on the existing platform.However,most of these platforms can't do accurate video retrieval based on knowledge point.How to help learners to quickly and accurately find specific knowledge point in the mass of resources,to meet their personalized learning needs,and to improve their learning efficiency are the core problems to be solved in this paper.Since the knowledge point of instructional video is usually the keyword of video content,the key to solve this problem is to realize automatic keyword extraction according to the teaching content of instructional video.Although there are many keyword extraction algorithms for text,there is no research on keyword extraction for instructional video with high colloquial degree and hierarchical knowledge.Therefore,there is an urgent need to propose a method of automatic keyword extraction based on MOOC video,The main contributions of this paper are as follows.(1)This paper analyzes and summarizes the language features of MOOC video,and constructs the test data set for keyword extraction and data preprocessing.Since there is no standard data set for instructional video at present,this paper constructs the test data set of MOOC by using the methods of video crawler and video phonetic writing.In order to ensure the accuracy of keyword extraction and obtain the candidate document set of keyword extraction,the text preprocessing of data set from Chinese word segmentation and text filtering is put forward.(2)Based on the classical Text Rank algorithm and the language features of MOOC video,a multi-feature fusion algorithm of Text Rank for keyword extraction based on MOOC video is proposed.Firstly,according to the language features of MOOC video,the feature extraction of candidate document set is carried out,and the features that affect the importance of words are obtained,such as word position,part of speech,domain feature,etc.Secondly,the weight distribution of word position,part of speech and domain feature in the text is determined by ordering relation algorithm.Thirdly,the word synthesis weight is used to determine the initial weight and the probability transfer matrix of the lexical nodes in the traditional Text Rank model,and the improved Text Rank algorithm is obtained.Finally,the final weight of each word is calculated by iteration of the algorithm,and some words with the largest weight are selected as keywords.In order to verify the performance of the algorithm,a simulation experiment is carried out on the test data set.The experimental results show that the proposed algorithm is feasible and effective.(3)Based on a multi-feature fusion algorithm of Text Rank for keyword extraction on MOOC video.A MOOC video keyword extraction system is designed and developed.The system has the advantages of simple interface,simple operation,humanization and expansibility of design,batch processing of data,simulation keyword extraction algorithm and calculation and evaluation index,etc.It has good engineering practicability.The research in this paper shows that keyword extraction on MOOC video can help learners quickly and accurately find specific knowledge point in mass MOOC video,meet their needs of personalized learning,and improve learning efficiency effectively.The automatic keyword extraction algorithm proposed in this paper is expected to be widely used in MOOC retrieval system.