| Online learning support service is one of the core contents of online learning,mainly in cognitive,emotional and social aspects,that is,academic and non-academic support services.The online learning forum is one of the main media for online learning support services.The content of the online learning forum post is generated by learners,which reflects the needs,emotion,social communication and other concerns of the learners during the learning process.Connetivism think that online learning is the process of building a connection between learners,creating a knowledge network,and exchanging knowledge.Online learning support services based on the online learning forum can also be seen as the process of creating the connetive learning network:(1)learning service supporters are looked as learners;(2)online learning support service is the process of learning,understanding,and sharing ideas about the "knowledge" generated by the learners’ posts.Faced with the massive connetive leaning network(online learning),the key issue of learning service supporters is how to quickly and effectively learn the knowledge generated and shared by learners(cognitive the concerns of learners).Based on the connectivism theory and the Social and cultural theory,the thesis employs the online learning forum data as the research object,with the effective use of the online learning forum posts content for online learning support service as the research goal,studies the techniques to use the information(such as the learners focus,emotion and social relationships,etc.)contained in the content of posts for online learning support service.And,the techniques will be applied to the learning support services.The detailed research work and innovation of this paper are as follows:(1)Online learning forum learning support service system construction.The thesis first defines the core concepts such as online learning support service.Then,analyze the characteristics of online learning forums and the components of online learning forum learning support service system;Next,design the framework of online learning forum learning support service system;Finally,analyze the technical support you need for your system based on the online learning support service provided.This set the scope for later research.(2)The construction and feature analysis of online learning forum dataset.At present,there are many studies on the contents of the online learning forum posts,but there is no online learning forum corpus data set for public research because it is difficult to obtain data due to privacy and other factors.In this paper,the posts of the public learning forum are collected and analyzed by using Regular Matching.And after desensitization,a real learning forum data set which contains more than 1.1 million posts,150 thousand learners is constructed.The data set is mainly composed of two parts:online degree education and online training.Then,the method extending the thesaurus of the word segmentation tool is used to solve the problem that the specific words can not be accurately segmented.And the sentence type judgment rules are proposed to solve the problem of sentence modal analysis.On this basis,the thesis deals with the word segmentation and the basic discourse feature analysis of the corpus.And it makes a comparative analysis of the features between online training and online degree education sub data set.The results of the analysis reflect the basic level of language proficiency of learners and learning support providers.This provides the basis for content mining and learning support service applications in the following chapters.(3)Learner focuses mining.The content of the online learning forum post reflects the difficulties and doubts faced by learners in the learning process,which is the true embodiment of the learning needs of online learners.Only understand the real needs of learners,learning support services can be done well.In order to grasp the focus of learners,a new dictionary of network education features(including 869 words)is constructed.The paper puts forward the content fusion method to optimize the content structure of thread,and puts forward the scheme of "Author-Topic"(AT)model +dictionary.By using this scheme,the authors and contents of optimized threads are mapped into the AT model,generate "Learner-Topic" model to discover the learners’ focus.The experimental results show that:(a)the topic is more clearly and accurately reflect the learners’ inner needs and interests,so that the goal of learning support services is accurate and clear;(b)with the increasing of topic number,the granularity of the topic meaning will gradually become more delicate,making leaning support services more detailed;(c)the expression effect of the topic depends on the features it contains,especially domain features,which makes the learning support service more professional and more purposeful;This requires efficient selection,filtering,and construction of domain dictionaries in experiments.The study of mining learners’ focus has laid the foundation for the application research that resource recommendation and discovery potential learning collaboration group of learning support services.(4)The analysis of learner’s affective support service based on Emotional Dictionary.Learning emotional support service is an important part of learning support services.The Semantic Orientation in the learner’s post not only reflects its learning needs and learning attitude,but also is an important factor in the learning performance during the individual learning process.Therefore,how to analyze the emotion contained in the learner’s post is an important issue.In order to solve this problem,firstly,the thesis puts forward the construction method of compound emotional dictionary,combining three open Chinese emotional dictionaries,such as CEVOD(Chinese emotional vocabulary ontology database)et al.,and expanding the network emotional words in it,and calculating emotional degree for without words,to generate a new high-quality compound emotional dictionary.In order to analyze the emotional intensity,we propose the emotional density function to decrease the influence of the length of the post text on the calculation of the emotional index.This function makes the emotion-rich post text to be moved to both ends of the emotional polarity.We further put forward the calculation methods of emotional index for post,sub forum,thread and learner.The experimental results show that these methods are effective for the online learning forum emotional analysis.This study provides support to observe the emotional changes of learners during the learning process and understand the"priority" of learners’ needs,and provides a research foundation for emergency problem and emotional counseling in learning support services.(5)An analysis of the social interaction of learners.The social interaction among online learners is an important social activity for learners to get rid of learning "loneliness" and enhance learning"collaboration".Mining and making use of this kind of sociality is one of the tasks of online learning support services.Therefore,this part of the study focuses on discovering the potential learning collaboration group(the groupment of learners)and the super-poster(the individuality of learners)in the online learning forum.During the process,we use the graph model to analyze the network interaction structure of the learners,propose to modify the centrality computing method of the traditional social network analysis,and increase the interaction intensity function to reflect the interaction strength.The two-stage clustering algorithm is proposed to discover the potential learning collaboration group in the online learning forum.The super-poster identification model is proposed to find the super-poster in the online learning forum.The experimental results show that the proposed algorithm and model are effective and feasible.This provides the basis for the study of recessive evaluation and learner profile in learning support services.(6)A study on the application of learning support service based on multi-tier architecture.Based on the above research,this part mainly studies the application research from the level of learning support service,and applies the research of learners’ focus discovery,emotion analysis and network social relation analysis to online learning support service.Firstly,the application architecture including basic data processing,information mining,learning support service application and application feedback four-layer structure is proposed.Secondly,under the framework of the support service system,the academic,non academic and emotional learning support service applications are proposed:resource recommendation,recessive evaluation,emergency problems and emotional counseling.Thirdly,according to the indicators of the individual learner such as the activity,the emotional index and the super-poster index and so on,the profile application of learner is proposed in the framework of learning support service system.Application analysis shows that these applications have a lot to help and promote the online learning support service.In summary,the thesis constructs the framework of online forum learning support service system,and defines the scope of the research at first.On this basis,the thesis constructs the online learning forum corpus data set and analyzes its basic discourse features.The paper further studies the techniques of using the data set to the learning support service,and put these techniques into application research,achieved good experimental results and application effect,proved the usefulness of this paper study for the online learning support service. |
PDF Full Text Request