Research On Error Data Detection And Repair Based On Online Learning Community Web Short-Text

With the development of Internet technology,online learning communities are increasingly appearing in our learning life,and more and more people are starting to learn independently through online learning communities.Learning analysis research refers to the analysis of data in the online learning community to help students make better use of the online learning community for personalized learning.However,short text data in the online learning community has some wrong information,such as true word errors,misunderstandings,etc.These erroneous data can lead to a large deviation between the research results of the learning analysis and the actual situation.Therefore,the research on error detection and repair of Web short text in online learning community has important theoretical research and application value.This thesis focuses on the detection and repair of Web short text errors.These errors mainly involve:Chinese real word errors and short text misunderstandings.Combining the relevant research work of previous researchers,the work of this paper is as follows:Firstly,based on the detection and repair of short text error data in the online learning community,this thesis proposes a short text error data detection and repair framework,which provides research ideas for the detection and repair of Web short text error data in the online learning community.Aiming at the problem of Chinese true word error and short text misunderstanding of short text in online learning community,this paper proposes a research framework of Chinese real word error detection and repair based on online learning community Web short text,and misunderstanding detection based on online learning community Web short text.The research framework is fixed.The former is used to guide the detection and repair of true words,and the latter is used in the research of short text misunderstanding detection and repair.Then,this thesis proposes a Chinese true word error detection and repair algorithm based on the online learning community Web short text,and a misunderstanding detection and repair algorithm based on the online learning community Web short text.The former is based on the improvement of three traditional algorithms.These three traditional word error detection and repair algorithms are based on n-gram,contextual context,and Chinese fixed matching.In this thesis,these three algorithms are improved and integrated into a unified algorithm to obtain the advantages of each algorithm,thus improving the accuracy of Chinese real word error detection and repair.The latter is an algorithm based on long short-term memory neural network(LSTM)and convolutional neural network(CNN),which converts text into a vector containing all the semantic information of the text itself through LSTM,and then uses the obtained vector as the CNN model.The input is trained to obtain a short text misinterpretation detection model.On the basis of this,combined with the output of the misinterpretation detection model,we propose a short text misunderstanding repair algorithm.Finally,the thesis carried out relevant experiments on the real data set,and used the three evaluation indicators of recall rate,accuracy and repair rate to evaluate the experimental results.The experimental results show that the Chinese real error detection and repair algorithm based on the online learning community Web short text and the misunderstanding detection and repair algorithm based on the online learning community web short text have good accuracy.