Zhang et al. Journal of Cloud Computing: Advances, Systems and Applications(2020) nal of Cloud Computing:Advances, Systems and ApplicationsRESEARCHOpen AccessA learning style classification approachbased on deep belief network for largescale online educationHao Zhang1,2*†, Tao Huang1,2†, Sanya Liu1,2†, Hao Yin3, Jia Li1,2, Huali Yang1,2 and Yu Xia1,2AbstractWith the rapidly growing demand for large-scale online education and the advent of big data, numerous researchworks have been performed to enhance learning quality in e-learning environments. Among these studies, adaptivelearning has become an increasingly important issue. The traditional classification approaches analyze only thesurface characteristics of students but fail to classify students accurately in terms of deep learning features.Meanwhile, these approaches are unable to analyze these high-dimensional learning behaviors in massive amountsof data. Hence, we propose a learning style classification approach based on the deep belief network (DBN) forlarge-scale online education to identify students’ learning styles and classify them. The first step is to build alearning style model and identify indicators of learning style based on the experiences of experts; then, relate theindicators to the different learning styles. We improve the DBN model and identify a student’s learning style byanalyzing each individual’s learning style features using the improved DBN. Finally, we verify the DBN result byconducting practical experiments on an actual educational dataset. The various learning styles are determined bysoliciting questionnaires from students based on the ILS theory by Felder and Soloman (1996) and the Readinessfor Education At a Distance Indicator. Then, we utilized those data to train our DBNLS model. The experimentalresults indicate that the proposed DBNLS method has better accuracy than do the traditional approaches.Keywords: Large-scale online education, Adaptive learning, Deep belief network, Learning style, High-dimensionalIntroductionWith the deepening integration of big data  and education, the learning revolution, represented by MOOCs, Khan academy  and Flipped classroom , hashad a strong impact on traditional forms of educationand has highlighted the importance of large-scale onlineeducation in reshaping education reshaping—that is,globalized resources, modularized supply, personalizedteaching and independent study can be implemented by* Correspondence: firstname.lastname@example.org†Hao Zhang, Tao Huang and Sanya Liu contributed equally to this work andshould be considered co-first authors1National Engineering Laboratory for Educational Big Data Central ChinaNormal University, Wuhan, China2National Engineering Research Center for E-Learning, Central China NormalUniversity, Wuhan, ChinaFull list of author information is available at the end of the articlerestructuring and process reengineering. Large-scale online education has made significant breakthroughs inteaching environments, content presentation, teachingmodes, and learning evaluations. Large-scale online education also innovates and revolutionizes study approachesand is an effective way to promote educational equity andimprove teaching quality. By offering new learning situations and technologies, the unique characteristics ofMOOCs have given rise to a new learning model, personalized learning [5, 6]. In a MOOC environment, learnerscan select and order customized courses according totheir own purposes and backgrounds. However, the materials in MOOCs are rather complicated; consequentlylearners can easily become trapped in “learning Trek” and“cognitive overload” situations  during the process ofreceiving information. Nevertheless, MOOCs provide the The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you giveappropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in this article are included in the article's Creative Commonslicence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commonslicence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtainpermission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Zhang et al. Journal of Cloud Computing: Advances, Systems and Applicationssame learning materials and learning activities for alllearners and ignore individual differences because they failto analyze individual learning behaviors. As a result,MOOCs cannot achieve the ability to teach students according to their aptitudes . By analyzing online learningbehaviors, we can accurately identify a user’s learningcharacteristics and recommend personalized resources tohelp them improve their quality of learning. At present, itis urgent to determine how to analyze and identify thelearning features of individual users by mining big data inadaptive learning [9, 10].In an online learning environment, individual differences between learners are clearly evident in terms oftime, learning duration, selected learning content, onlineinteractions, etc. Some learners prefer to complete learning tasks during the daytime, while others prefer learning at night. Some learners participate in discussions orgroup activities on discussion forums or though othersocial networking applications, while others prefer quietstudy. Among all the individual characteristics, learningstyle is an important factor that affects learners’ individual differences. In other words, different learners havedifferent tendencies in terms of learning style .To date, researchers have made important contributions regarding how to apply learning styles to onlinelearning, especially in the field of learning style identification and prediction [12–15]. Most of these studiesmainly collected and recorded real behavior data left bylearners during the network learning process to build aset of learners’ network learning behaviors, and useddata mining algorithms, neural networks or simple calculation rules to automatically detect learning; thussome research results have been obtained. However, thecentral problem is that these studies were not based onlarge-scale online learning platforms; consequently, theydo not meet the demands of modern online learningplatforms for effective learning style detection based ona large number of complex network learner behaviors.Therefore, it is urgent to solve problems such as howto identify learners’ learning styles, how to guide individual learners to construct learning objectives and plans,and how to recommend specific learning resources thatreflect an individual’s needs and abilities. Identifyinglearning styles through learners’ online behavior andclassifying them correctly plays an important role inrealizing adaptive learning. In ta MOOC environment,the following three important problems must be solvedto accurately identify and classify learning styles:1) What type of learning style model is suitable inMOOC learning environments?2) What is the relationship between learning behaviorand learning styles in MOOC learningenvironments?(2020) 9:26Page 2 of 173) What classification method can be used toovercome the problem of inaccurate classificationdue to the high dimensionality of learning behaviordata in MOOC environments?Through intensive study of the relevant literature concerning learning styles, we found that the learning stylemodel proposed by Felder-Silverman is suitable forMOOC online learning environments. Meanwhile, theonline learning relationship between behavior characteristics and learning styles is complex. Through our observations, together with the results obtained from experts,we found that differences in student behavior can be described using the following four dimensions: informationprocessing, perception, input, and understanding. Therefore, the characteristic indicators should be developedbased on these four learning behavior dimensions toform the data mining dimension.We propose a learning style classification based on aDBN for MOOCs, called DBNLS. In this method, the inherent characteristics of learners—their learning styles—are adopted as the classification criteria. First, we summarized and analyzed individual learner differences and preferences to build a learning style model suitable forMOOC learning environments. Then, we determinedwhich learning-habit indicators should be used based onexpert experience and linked the indicators to learningstyles within individual sessions. The deep learning modelDBN was used to learn the high-dimensional learningstyle features and model learning styles to classify studentsaccurately. Finally, we collected network learning behaviordata by analyzing the weblogs left during online learningsessions on StarC , a MOOC platform used at CentralChina Normal University. Meanwhile, we also conductedoffline empirical research and collected learning stylequestionnaire data, which we used as training samples totrain the DBN model. The trained DBNLS model was applied to classify students’ learning styles. The results showthat the method proposed in this paper is superior to thetraditional methods.The main contributions of this paper are as follows:1) Learners’ intrinsic characteristics and learning stylesare introduced as an important standard for learnerclassification. Simultaneously, the explicit attributesof the network behavior are effectively mapped tothe intrinsic characteristic learning style indicators.These network behavior indicators serve asimportant DBN inputs for classifying students’learning styles.2) We introduce social interaction factors into thelearning style model, with the view of including thenetwork learning environment characteristics andthe learner interactions in the learning platform.
Zhang et al. Journal of Cloud Computing: Advances, Systems and ApplicationsThe learning style model established in this paperfocuses not only on the static learning resourcesavailable to students but also on how they interactwith others. For example, some students tend tocomplete tasks through communication anddiscussion, while others tend to think and workindependently.3) We introduce a deep-learning algorithm into learning classification in the field of education. This approach effectively overcomes the problems of thesharp rise in computational complexity resultingfrom high-dimensional data in traditional classification methods and data overfitting conditions.4) The research was conducted using a practicalonline learning activity. We collected anddistinguished both online and offline data. Theoffline data served as training data used to train themodel, which subsequently was used to classifystudents’ online learning behaviors. The resultsshow that the proposed mechanism is considerablymore accurate than is the traditional classificationmodel.The rest of this paper is organized as follows: Relatedwork and background introduces the related works andprovides some background concerning learning styles,the restricted Boltzmann machine (RBM) learningmodel, and BP. We then present the learning stylemodel and the classification model in Learning stylemodel in MOOCs. We present our DBNLS algorithmand its training model in Learning style detection basedon deep belief neural networks. The details of the experimental evaluation are described in Section 5, andthis work is concluded in Conclusions.Related work and backgroundResearch on learning style theoryThe concept of “learning style”, first defined by HerbertThelen, has since evolved into dozens of learning styletheories and has been put into practice in the field ofeducation. With the teaching methods described in“Teaching Students in Accordance with Their Aptitude”and “Learner Centered”, an increasing number ofscholars have shifted their focus to the learning style oflearners, which is hoped to be fully considered in theMOOC design process. In recent years, the rapid development of online educational tools such as MOOCs hasinspired scholars to consider how to reflect differentlearning styles in online education so that appropriatematerials and methods could be suggested accordinglyto help improve learning efficiency [11–15, 17].Theories on learning style have become relatively mature after a long research history. Many scholars haveproposed sophisticated learning style models. According(2020) 9:26Page 3 of 17to Curry’s learning style model, all learning styles can beclassified into three levels, namely, “teaching preference”the outer level, “information processing mode” the middle level, and “cognitive style” the innermost level .Learning style models of this type include Kolb ,Honey and Mumford , Dunn  and FelderSilverman’s learning style models . In addition, othermodels such as cognitive styles, VARK, and Keefe’slearning style model, propose different definitions oflearning style.Dunn’s learning style model is the representative theory of the “onion model” at the outer level. Dunn wasmainly concerned about the stimuli that influence learning activities . These stimuli are related to the learning environment, social environment, physiologicalfactors, psychological factors and emotional factors.However, all these stimuli are highly unstable and vulnerable as can easily be observed. In contrast, Kolb wasinterested in the learning process. He suggested thateach learning process goes through four self-relatedphases. Learners exhibit different preferences towardsthese phases . Based on the learning models in [19,22], and , Felder-Silverman provided a brand-newlearning style model  that focuses on learners’ individual cognitive characteristics and provides a comprehensive description of learning styles by combininginformation processing, information cognition, information input and information understanding.Felder also designed the Solomon Learning Style Scalebased on the learning style model, which provides agood method for measuring learning styles. As a result,Felder-Silverman’s model is not only widely used inpractice but also suitable for a web-based learning environment. Moreover, the Solomon Scale enjoys fairly goodreliability and validity and has become popular in theeducational field [23, 24]. Although scholars have provided different definitions, they all include the threemain characteristics of learning styles. First, learningstyle varies among individuals, which means that different learners tend to have different learning style preferences. Second, learning style formation is affected bystimuli from both the outside environment and the innerself, such as cultural discrepancies, family factors, educational factors, and physiological factors. Third, learningstyle affects learning behavior. Learners with differentlearning styles show differences in learning strategiesand learning habits.Because the measurement approach biased toward essential characteristics (cognitive characteristics) by FelderSilverman fits web learning, and Felder-Silverman’s modelhas been proved through numerous experiments to havegained a high frequency in practice, as shown in Table 1,this paper chooses Felder-Silverman as the base learningstyle model . However, characteristics such as social
Zhang et al. Journal of Cloud Computing: Advances, Systems and ApplicationsTable 1 The frequency of typical learning model uses in realenvironmentsLearning style model (n 70)numberpercentFelder-Silverman3550Cognitive styles1217.14Kolb68.57VARK57.14Honey and Mumford45.71Other34.29Not specified57.14interactions in new learning environments, such asMOOCs, were not considered in Felder-Silverman’smodel even though they should be considered in weblearning. Therefore, although based on the FelderSilverman model, this paper will expands the social interaction dimension to make it more suitable for a MOOCteaching environment.Learning style identificationThe traditional questionnaire approach to measuringlearning style does not fit well in a MOOC teaching environment—primarily because factors such as the subjective consciousness of the interviewees, failure tounderstand questions, and learning preferences measured at a specific point would negatively affect the accuracy of the results. Therefore, an increasing number ofscholars both domestically and internationally have instead been studying learning styles through automaticdetection methods [26–30]. An automatic detectionmethod is a way to detect learners’ learning styles automatically by collecting real data recorded about learnersby web learning platforms and adapting data mining,neural networks or simple calculation rules to the set oflearning behaviors that arise in in web learning contexts[31–35].The University of Vienna analyzed learning platformdata and web logs formed during learning platform sessions to recognize learners’ learning behaviors . In, the authors predicted students’ learning styles byanalyzing log data using BP neural networks, while used a Bayesian network approach to recognize learningstyles of students attending artificial intelligence onlinecourses and discovered a significant disparity in the accuracy of predicting learning styles from different dimensions  combined decision tree and hiddenMarkov models to evaluate learning behaviors and solvedifficulties in sequence data to more accurately analyzeof sequential and comprehensive learning styles fromthe understanding dimension. A contrast test was conducted in  on the teaching efficiency of adaptivelearning styles. The results show that the students who(2020) 9:26Page 4 of 17attended adaptive courses acquired a higher learning efficiency and performed better on examinations. The authorsof  reported that standard achievement assessmentscould not only assess students’ learning abilities but also detect individual learning characteristics and predict the results . analyzed and processed the web pages visited byweb-based learners to study learning styles, and  recorded web-based learners’ learning demands and activitiesand explored their individualized features to study learningperformance assessments.Both the traditional learning algorithms and the backpropagation with adaptive learning rate (BPAL) networkalgorithm have two weak points. The first is that the rawdata cannot include too many properties; the greater thenumber of properties are, the greater the difficulty iswhen composing a corresponding vector. When executing the classification algorithm, the computational complexity increases exponentially as the vector lengthincreases. Second, the mapping relations between theproperties of the raw data and learning styles cannot betoo complex. Thus, these algorithms are unsuitable forcomplex mapping relations. The conventional methodsfail to analyze and process the complex relations between the behavior data of web learning and learningstyles. However, deep learning is a way to extract characteristics from vectors in a step-by-step manner, allowingmore useful features to be studied by building a machinelearning model with multiple hidden layers and enormous amounts of training data to promote the classification and prediction accuracy.Deep belief networksDeep belief networks (DBNs) were first proposed byHinton et al. in 2006 . A DBN analyzes the potentialfeatures of texts, images and voice by constructing amultilayer neural network model . The training dataproceeds through the network layer by layer, and eachlayer extracts more advanced features than the previouslayers. Deep lea
RESEARCH Open Access A learning style classification approach based on deep belief network for large-scale online education Hao Zhang1,2*†, Tao Huang1,2†, Sanya Liu1,2†, Hao Yin3, Jia Li1,2, Huali Yang1,2 and Yu Xia1,2 Abstract With the rapidly growing demand for large-scale onlin
green, Chicago-style/verde de estilo Chicago 11.04.07 Center for Urban Ecology, University of Illinois@Chicago . 0 200000 400000 600000 800000 1000000 1200000 1400000 "Chicago-style" "New York style" "L A -style" infomal comparisons/ Chicago style "Chicago style" "New York style" "LA style"
classification has its own merits and demerits, but for the purpose of study the drugs are classified in the following different ways: Alphabetical classification Morphological classification Taxonomical classification Pharmacological classification Chemical classification
Jul 08, 2018 · Song: Grease (Is the Word), page 2 of the featured songbook Style: 8-Beat Rock Tempo: 110 /- Setup: 8-Beat Rock style: Intro Normal Style Setup #9, Song Normal Style Setup 0 (zero), 2, 5, and 9. Pretty Woman style: Intro Vintage Style Setup #10, Song Vintage Style Setup 0 (zero), 1, 6, and 10. Poerfmanr ce: Use a pencil to mark the music. Play to 1st ending and press Setup #2 (#1 for .
25 lotus petals a collection of martial arts styles for use with white wolf publishing’s exalted rpg terrestrial blossoming cherry tree style 1 brave dragon taunting style 3 jade fan style 6 rippling quicksilver style 8 celestial black mourner style 10 argent scorpion of opposition style 12 crimson temple style 21 fea
AMBASSADOR TERRACE EAST 1374 SQ.FT. AMBASSADOR TERRACE WEST BALCONY 1031 SQ.FT. DN r 1 2" r 1 2" r 1 2" ROOM East Banquet Style Boardroom Style Reception Style Classroom Style Rounds Theatre Style Ushape Style Hollow Style Square Feet Wheelchair Access 70 25 110 45 120 30 25 1374 Yes LEGEND T X'-X" Single Light Switch 2-gang Light Switch 3-gang .
used in these Classification Rules have the meaning given to them in the Glossary to these Classification Rules. 1.12 References to a 'sport' in these Classification Rules refer to both a sport and an individual discipline within a sport. 1.13 The Appendices to these Classification Rules are part of these Classification Rules
Unit: I wanna play in a band Style: Rock Unit: Zootime Style: Reggae Unit: Friendship Song Style: Pop Unit: Reflect, Rewind and Replay Unit: Let your spirit fly Style: R&B, Soul, Musicals Unit: Glockenspiel Stage 1 Style: Learning basic instrumental skills by playing tunes in varying styles Unit: Three Little Birds Style: Reggae Unit: The .
Agile methods in SWEP Scrum (mainly) XP Head First Software Development Process The Scrum process follows the agile manifesto is intended for groups of 7 consists of simple rules and is thus easy to learn 15.04.2012 Andreas Schroeder 9