Adaptivity In E-learning Systems Loreta Leka Alda Kika Silvana Greca Informatics Department Informatics Department Informatics Department Faculty of Natural Sciences Faculty of Natural Sciences Faculty of Natural Sciences University of Tirana University of Tirana University of Tirana loreta.leka@fshn.edu.al alda.kika@fshn.edu.al silvana.greca@fshn.edu.al interactive system that adapts its behaviour to indi- vidual users on the basis of processes of user model Abstract acquisition and application that involve some form of learning, inference, or decision making [Jameson 2009]. This paper aims to give a short review of adap- Research has shown that the application of adapta- tivity in e-learning systems and the work done tion or personalization can provide a better learning in this field. The review is mainly focused on environment since learners perceive and process infor- the different parameters we can use to make mation in very different ways. So, the adaptive edu- an e-learning system adaptive. Some adaptive cational systems are an alternative to the traditional e-learning systems are shortly described, high- teaching; they can be considered to be the next gen- lighting the student characteristics the system eration of e-learning [Per08]. adapts to. Deciding the parameter or param- Before designing an adaptive e-learning system, one eters the system will adapt to, is the first step of the main challenges is to identify which learner in designing an adaptive e-learning system, needs or characteristics should be made adaptive. Re- which is the final goal of the future work that searchers have suggested different approaches, and can be done, discussed in the conclusions sec- some adaptive e-learning systems are designed. In tion. The final system will be used in educa- this paper, a short review of adaptivity parameters is tion, as a helpful system to come to student described, and then some adaptive systems are men- needs, and increase their learning performance tioned, highlighting the adaptivity parameter they use. and motivation. Adaptive models are analyzed further, being a central part of designing an adaptive system. The paper closes 1 Introduction with a conclusion and future work. Nowadays, e-learning is a widely discussed topic, and many e-learning systems have been developed so far. 2 Adaptivity parameters However, traditional e-learning systems tend to ne- glect the diversity of learners, their abilities, their A system that automatically adapts to the student, knowledge and skills, and the learning context [J14]. based on its assumption about the student, is referred The lack of adaptive learning environments or an en- to as an adaptive system [itk15]. In other words, the vironment with adaptive features is partly due to the system cannot be called adaptive if it is not flexible to concept one-size-fits-all.Very often, e-learning courses specific students needs. This leads to the fact, that de- have a problem of universal size as the same static con- ciding which student feature or characteristic to make tent is presented to all students. Adaptive e-learning adaptive, in order to come more closely to students systems are the new trend in e-learning systems.Their needs, while building this adaptive e-learning system, ultimate goal is to personalize learning material and is one of the key decisions, and one of the main fac- their sequences to match the needs of an individual tors to indicate its success. In the past decade, various learner as closely as possible. These systems integrate adaptive learning systems have been developed based learner characteristics such as learning style, affective on different parameters that represent the characteris- state and knowledge level to provide personalized ser- tics or preferences of students as well as the attributes vices and recommend relevant instructional material of learning content [Wan11]. Based on a review done [Bru01]. Jameson describes an adaptive system as an with various systems built, we are going to overview some of the main parameters used. 2.1 Adaptation To Student Knowledge One common example of adaptation in an e-learning system is the adaptation of the learning materials, con- tent presentation according the knowledge of the stu- dent in the subject area. The main idea is that for an advanced student, the system can provide a brief summary of the material and hyperlinks to the more detailed description of it. In the case of a learner who has little knowledge on the field, the system can pro- vide more detailed information in a smooth logical flow [Puu05]. One system that uses this method is ELM- ART. It is an adaptive e-learning system used to learn Lisp programming. It was one of the first and most influential adaptive e-learning systems [P01], so much Figure 1: Learning styles adaptive systems [P99] [P04] so that its last version, of 2001, remains in use until [Sch08] [iWe03] [Bac11] [K02] today to learn Lisp programming. It adapts learning material according to each learners knowledge level. nisms that allow humans to acquire and recognize [Bru15]. SQL-Tutor, another example, is an intelli- pieces of information, to convert them into representa- gent tutoring system that personalizes SQL learning tions, then into knowledge, and finally to use them for concepts according to the individuals knowledge level. the generation of simple to complex behaviors [Sot08]. It selects some questions in basis of learners model, There are four cognitive abilities: working memory ca- then it adapts the model based on the answers valid- pacity, inductive reasoning ability, associative learning ity [Hau04]. ability, and speed of information processing. Research has suggested that cognitive abilities along with learn- 2.2 Adaptation To Learning Styles ing styles are very important factors for learning effi- This method of adaptation is based on the idea that ciency, so it should be considered in designing adaptive a student can learn more efficiently given the mate- systems. AES-CS system is an intelligent system that rial according to his learning style. Different people recommends relevant learning material based on the have different learning styles. Researchers have pro- Witkin model of cognitive style: field dependence and posed different learning style theories or models. Some field independence. of them are: The Felder-SilverMan model, the Dunn and Dunn Model, the Kolb Model, the Witkin Model 2.4 Adaptation To Learning Behavior And etc. The model which has been recognized from many Motivation researchers as highly suitable for adaptive e-learning Tracing learners behavior in real time is a quite chal- systems is the Felder-Silverman model. This model lenging task. In her work, [Conati, 2002] address the categorizes ways students process information based problem of how an interactive system can monitor the on these groups: sensory and intuitive, visual and au- users emotional state using multiple direct indicators ditory, inductive and deductive, reflective and active, of emotional arousal. Detection of users body expres- generally and sequential. Based on each group, the sions requires special sensors. The system was applied appropriate teaching method is used for each partic- on computer-based educational games instead of more ular student. There have been a number of adaptive traditional computer-based tutors, as the former tend e-learning systems built, using this approach, as de- to generate a much higher level of students emotional scribed in the figure 1 engagement. It has been argued that if a learner has a strong Another approach used is real time eye tracking. In affinity for a particular learning style, the learning ma- [Gutl et al., 2005] the authors introduced the Adaptive terial and strategies should match this style to enhance e-Learning with Eye-Tracking System, a system that learning [K02]. utilizes a monitor mounted camera that records the eye of the participant and trace the gaze in a scene 2.3 Adaptation To Cognitive Abilities through imaging algorithms. Real- time information of According to [Riding and Rayner, 1998] Cognitive the precise position of gaze and of pupil diameter can Style (CS) refers to an individuals method of pro- be used for assessing users interest, attention, tiredness cessing information. Cognitive abilities are mecha- etc [Geo10]. 3 Adapting To Multiple Methods There is also a number of adaptive e-learning systems that integrate both learning style and knowledge level as learner characteristics that drive adaptation:For ex- ample, MASPLANG is one of the pioneers, combin- ing both learning style based on the Felder-Silverman model and knowledge level to adapt learning material related to a computer networking course. One recent example of a successful system is Protus, an adaptive e-learning system based on learning style and knowl- edge level that recommends relevant learning material for teaching the Java programming language [Mil11]. Although there are different adaptation techniques for e-learning systems, the common idea for them all is that each student must learn the way he prefers, the adaptation must be done frequently, no one should Figure 2: AHAM Model [Bra99] continue to learn something that is completely learned successfully, and each student must be presented with main. The content of domain models are those that different information of a certain subject until he has are adapted to the different needs of learners in adap- learned it successfully. tive e-learning systems. Learning objects are usually organized and annotated using metadata in order to 4 Adaptive Models describe, sequence, store and manipulate them. For example, Sun, Joy and Griffiths have proposed a novel Adaptive models represent an important research area. mechanism to categorize learning objects according to They can be used to form the design and develop- the Felder-Silverman learning style model in order to ment of adaptive e-learning systems, taking into ac- dynamically provide relevant learning objects to each count their main components. Mainly adaptive models learner according to their learning style preferences answer these three questions: what can we adapt (do- [Gri07]. They proposed a multi-agent system which main model), to what we can adapt (student model), stores each students current learning style and the and how can we adapt (adaptation model). One popu- style attributes of each learning object. Initially, the lar approach is the Dexter Hypertext Reference Model, student style is set based on Felder-Silverman ques- which can be used as a logical foundation for designing tionnaire to determine students style. Each learning and comparing different adaptive systems. The model object is also categorized based on the learning styles. consists of three layers including a run-time layer, a The system searches the repository of learning objects, storage layer and a within-components layer.The stor- and fetches the appropriate learning object based on age layer refers to how contents are connected and the student learning style. The Learning Object Agent stored in a database. The run-time layer deals with is responsible to provides relevant learning objects for the representation of user interaction and hypertext. students with different learning styles. The within components layer deals with the content and structure of components within a hypertext net- 4.2 Student model work. The Dexter model has influenced the design of many interactive web-based systems [Hal94]. An ex- In the area of the Web systems the user models have tension of the Dexter model was developed to support the task to manipulate information that refer to the adaptively, called the Adaptive Hypermedia Applica- knowledge of a user in a specific domain, to his/her tion Model (AHAM) [Bra99]. AHAM enhanced the personality, his/her preferences, or to any other in- storage layer of the Dexter model by adding three sub- formation that can be useful in the customization of models including a domain model, a user model and an application. The student model stores information an adaptation model. that is specific to each individual learner: it concerns how and what the student learns or his/her errors, and the student model plays a main role in planning the 4.1 Domain model training path, supplying information to the pedagog- A domain model is an abstract representation of part ical module of the system. This component provides of the real world. It is composed of a set of do- a pattern of the educational process, using the stu- main knowledge elements and is the result of captur- dent model in order to decide the instruction method ing and structuring knowledge related to a specific do- that reflects the different needs of each student [Lic04]. Figure 3 provides an abstract representation of the stu- 4.3 Adaptation model dent model and its content. An adaptation model bridges the gap between the learner model and the domain model by matching relevant learning material, or sequence of objects, to the needs and characteristics of an individual learner [Lic04]. The adaptation model is strongly related to the student model. According to student model, it adapts and recommends relevant learning material. Based on the design of the system, the adaptation model can adapt using short memory cycle, or long memory cycle. In the first case, the adaptation is done based on recent information about the user; for example after completing a test. In the second case, the system takes into account historical information in addition to recent one, to make the adaptation. The adaptation model can incorporate different adaptive methods and techniques to support adaptation. They can be included in these categories: adaptive naviga- Figure 3: An abstract represantation of the student tion, adaptive content, adaptive presentation. model [Alshamari 2016] 4.3.1 Adaptive Navigation User (learner) modeling involves different stages Adaptive navigation recommends selective learning such as data elicitation, model representation and paths or curriculum sequencing. Other examples in- maintenance. Data elicitation is usually based on clude link generation, direct guidance and link hiding. explicit methods via user generated feedback (such as questionnaires, like/dislike and rating) or implicit 4.3.2 Adaptive Content methods, which consider system generated feedback The idea behind this technique is that the system can (such as mouse movements, time spent and page vis- choose from the learning material the most appropri- its). Although explicit methods are considered more ate fragment of content based on the user model. reliable and more accurate, learners may be reluc- tant to provide explicit feedback. In contrast, im- 4.3.3 Adaptive presentation plicit methods allow learners to focus entirely on their Adaptive presentation is related to zooming, scal- main task. A large amount of data can be captured ing and layout-changing techniques. Another classic through an implicit method [Lic04]. In many learn- adaptive navigation technique is personalized learn- ing systems, learners are allowed to interact and up- ing paths. This generates different learning paths for date their own learning model. Students model is up- learners based on their preferences, learning style or dated either on the basis of test performance or a stu- knowledge level [J14]. dent can himself update by marking concepts known to him. Learner modeling is done on two different time scales: long term and short term modeling . The 5 Conclusions long term modeling attempts to model those aspects Designing an adaptive e-learning system is still an up of a learner that are not expected to change too dy- to date topic. Although researchers have proposed dif- namically. The short-term modeling is also being per- ferent models, they are mainly experimental, and very formed in two ways: indirectly and directly. Indirect few have become commercial or really used. Another short term modeling includes counting the number of challenge is integrating these systems with our educa- times a learner reviews a learning object, measuring tional system, especially in universities. Students are the total time taken to complete the topic. Direct faced with a large amount of material to study, and short-term modeling is carried out by assessment on often they dont know how to filter it, and lack motiva- questionnaires that evaluates the learner performance tion for studying. An adaptive e-learning system can as a skill level. More complicated techniques, such be a helpful, being in the role of the personal tutor as Bayesian belief networks can be effectively used to for them. They can also be aware of their knowledge construct student models. Several researchers have ex- or expertise of a field, and have a clear idea about plored the use of Bayesian belief networks to represent their personal level, every time. There are many chal- student models [Gre04]. lenges for this idea, the right methods and strategies to meet student needs should be applied. This paper [Gri07] Sun Joy Griffiths. “The Use of Learning Ob- has given a shortly review of these adaptive systems, jects and Learning Styles in a Multi-Agent ways we can make it adaptive, and main components Education System ”. In: (2007). of an adaptive system. In the future, the aim is to de- [Per08] Elena Verdu Luisa M. Reguera Maria Je- sign a convenient system to be applied at universities, sua Verdu Juan Pablo De Castro Maria nge- which will facilitate students and pedagogues work. les Perez. “An analysis of the Research on Adaptive Learning:The Next Generation of References e-Learning”. In: (2008). [Hal94] Halasz. “The Dexter Hypertext Reference [Sch08] Schiaffino. “eTeacher: Providing personal- Model”. In: (1994). ized assistance to e-learning students”. In: (2008). [Bra99] Hongjing Wu Geert-Jan Houben Paul De Bra. “AHAM: A Reference Model to Sup- [Sot08] Dimitrios Georgiou Sotirios Botsios. “Re- port Adaptive Hypermedia Authoring”. In: cent Adaptive e-learning contributions to- (1999). wards a standard ready architecture”. In: (2008). [P99] Carver C. A Howard R. A. Lane W. DWe- ber G. Brusilovsky P. “Addressing differ- [Geo10] Dimitrios Georgiou. “Using Standards for ent learning styles through course hyperme- Adaptive Learning Objects Retrieval”. In: dia.” In: IEEE Transactions on Education (2010). (1999). [Bac11] Bachari. “E-learning personalization based [Bru01] Peter Brusilovsky. “Adaptive Hypermedia”. on dynamic learners preference”. In: (2011). In: (2001). [Mil11] Klasnja Milicevic. “E-Learning personaliza- [P01] Weber G Brusilovsky P. “ELM-ART: An tion based on hybrid recommendation strat- adaptive versatile system for Web based in- egy and learning style identification”. In: struction. International Journal of Artificial (2011). Intelligence in Education”. In: (2001). [Wan11] S. L Wu C. Y Wang. “Application of [K02] Felder R. M. Silverman L. K. “Learning and context-aware and personalized recommen- teaching styles in engineering education”. dation to implement an adaptive ubiquitous In: (2002). url: http://www.ncsu.edu/ learning system”. In: (2011). felderpublic/%20Papers/LS-1988.pdf. [J14] Alshammari M Anane R Robert J. “Adap- [iWe03] Wolf iWeaver. “Towards ’Learning Style’- tivity in E-learning systems”. In: Proceed- based e-Learning”. In: (2003). ings - 2014 8th International Conference on Complex, Intelligent and Software Intensive [Gre04] J. D. Zapata-Rivera J. E. Greer. “Inter- Systems, CISIS 2014 (2014), pp. 79–86. acting with Inspectable Bayesian Student Models ”. In: International Journal of Arti- [Bru15] Gerhard Weber1 Peter Brusilovsky2. “An ficial Intelligence in Education 14(2) (2004). Interactive and Intelligent Web-Based Elec- tronic Textbook”. In: (2015). [Hau04] Antonija Mitrovic Kurt Hausler. “An Intel- ligent SQL Tutor on the Web”. In: (2004). [itk15] Ani Grubii Slavomir Stankov Branko itko. “Adaptive Courseware: A Literature Re- [Lic04] Esposito Floriana Giovanni Semerano Ori- view ”. In: (2015). ana Licheli. “Discovering Student Models in e-learning Systems”. In: Journal of Univer- sal Computer Science 10.1 (2004), pp. 45– 57. [P04] Paredes P. Rodrguez P. “A mixed approach to modelling learning styles in adaptive edu- cational hypermedia. Advanced Technology for Learning”. In: (2004). [Puu05] Ekaterina Vasilyeva Mykola Pechenizkiy Seppo Puuronen. “Knowledge Manage- ment Challenges in Web-Based Adaptive e- Learning Systems”. In: (2005).