2.2.1. Psycho-pedagogical level

At this level, the pedagogical and the learning view of the process are not defined. The educational process is viewed as a psycho-pedagogical change that occurs by using the educational process meta model. The latter adopts various subjects, which are defined by different theories of learning paradigm.

2.2.2. Didactic level

At this level, the process is specified by adopting the interaction between the pedagogical and the learning view. By considering the latter, the process model is used to achieve learning outcomes by considering the constraints of the learning domain. For the pedagogical view, the process is viewed as a pedagogical goal achieved via defining a process model based on pedagogical content.

2.2.3. Situational level

This is the level of the instantiation of process models by considering the different characteristics of the learning/teaching situation in order to reach the desired learning level through the use of different learning objects. In fact, the situation characteristics are the objective, the different tasks and the different available resources.

2.2.4. Online level

This is the level of the execution of different learning and pedagogical actions that are supported by different learning systems. This execution is achieved by adopting different electronic media in order to use the different learning objects.

2.3.1. Didactic domain world

This is the world of processes taking into account the notion of the process and its nature (Bayounes et al. 2012). The didactic domain world contains the knowledge of the domain about which the proposed process has to provide learning (Bayounes et al. 2012). This world specifies the process nature by reviewing the domain dimensions. The nature view defines three facets, namely pedagogy, learning and process.

The pedagogy facet presents three attributes: the pedagogical orientation, the pedagogical method and the correlation. The different orientations are teacher-directed, learner-directed and teacher–learner negotiated. The pedagogical method specifies the activities of the learning process. It can be classified as direct instruction, indirect instruction, interactive instruction, experiential learning and independent study. The third attribute tests the correlation between the pedagogical and learning processes.

The learning facet presents three attributes: the performance, the learning mode and the learning outcome. The different performance types are remember, use and find (Merrill 1983). These performances are accomplished by three different learning modes, namely accretion, structuring and tuning (Rumelhart and Noorman 1978).

2.3.2. Instructional design world

This world deals with the representation of processes by adopting predefined models. It focuses on the description view by respecting the notation conditions and the constraints of representation level. This world includes the notation and the level of process description.

In the notation facet, four attributes are found: the type, the form, the content and the models. The two notation types are standard or proprietary. These types are used by adopting three major forms namely informal, semi-formal and formal. According to the component display theory (Merrill 1983), these forms present four major types of content. These types are fact, concept, procedure and principle. These contents are used to instantiate the domain model, the learner model and the pedagogical model.

Furthermore, the level facet presents three attributes: the granularity, the modularization and the coverage. The different levels of granularity are course, sequence of activities and simple activity. Three types of modularization, namely primitive, generic and aggregation, define these levels (Kumar and Ahuja 2020). In fact, the modularization attribute is used to capture the laws governing the learning process construction by using process modules (Bayounes et al. 2012). This process has been defined differently in different coverage (Bayounes et al. 2014):

The activity: the process is defined as a related set of activities conducted for the specific purpose of product definition.

The product: the process is a series of activities that cause successive product transformations to reach the desired product.

The decision: the process is defined as a set of related decisions conducted for the specific purpose of product definition.

The context: the process is a sequence of contexts causing successive product transformations under the influence of a decision taken in a context.

2.3.3. Learning environment world

This world deals with the entities and activities that arise as part of the engineering process itself (Bayounes et al. 2012). It focuses on the design view by describing the implementation of process representation. The design view includes five facets, namely context, construction, optimization, guidance and adaptation.

The first facet defines the type and the form of context. The construction facet deals with four issues: the approaches, the methods, the tools and the techniques of adaptive learning process construction. In a manner analogous to the adaptation spectrum (Patel and Kinshuk 1997), we can organize construction approaches in a spectrum ranging from “low” to “high” flexibility. The construction approach attribute is defined as the approach: ENUM {Rigid, Contingency, On-The-Fly}. These approaches are adopted by using the instantiation, the assembly and the ad hoc method. These methods apply three major techniques (curriculum sequencing, intelligent solution analysis, problem-solving support).

In the optimization facet, four attributes are found: the model, the method, the technique and the parameters of optimization. The different models are map-based, network-based and tree-based. Three types of methods, namely exact, heuristic and meta-heuristic, apply these models. These methods use rule-based and case-based techniques. The achieved educational intentions can be used to implement these optimization techniques.

In addition, the guidance facet defines four attributes: the type, the method, the outcomes and the parameters of guidance. The guidance types are strict and flexible. These types make use of three major methods, namely expository, collaborative and discovery. The outcomes of these methods are activity, resource and intention. To this end, the guidance methods adopt three parameters, namely learning style, cognitive state and teaching style.

Finally, the fourth facet introduces the adaptation by defining five attributes, which are the dimension, the position, the method, the technique and the parameters of adaptation. The three dimensions of adaptation are content, structure and presentation. The position is identified by adopting the adaptation spectrum (Patel and Kinshuk 1997). Each position is satisfied by applying four major methods, namely macro adaptive, aptitude treatment interaction, micro adaptive and constructivist-collaborative (Garcia-barrios et al. 2004). These methods are implemented by four major techniques, which are adaptive interaction, adaptive course delivery, content discovery and assembly and adaptive collaboration support (Paramythis and Loidl-Reisinger 2004). These techniques adopt four major parameters, specifically learning goal, learning history, prior knowledge and system information.

2.3.4. Learning situation world

This world supports the scenario view by examining the reason and the rationale of learning process engineering (Bayounes et al. 2012). It describes the organizational environment of the educational process by indicating the purpose and policy process management. The purpose facet includes two attributes: process and learning. In fact, the construction approaches have been designed for different purposes and try to describe the learning process in different attitudes: descriptive, prescriptive and explanatory. The learning purpose defines the level acquired by constructing the learning process. According to Bloom’s taxonomy (Forehand 2012), the different levels are knowledge, comprehension, application, analysis, synthesis and evaluation.

The second attribute identifies three policies of process management, particularly evolving, reuse and assessment. This attribute supports the validation of construction process quality. Moreover, since the learning situations change and evolve over time, it is essential that the learning process construction supports these evolutions (Bayounes et al. 2012). As with any process development, the reuse and the assessment, which can occur at any stage of learning process construction and at any level of abstraction and may involve any element of design and/or implementation, are outstanding.

2.4.1. Study scope

Several intelligent tutoring systems have been reported in the literature. Before applying our framework of educational process construction to these systems, Table 2.1 presents a brief description of selected systems. Indeed, this selection is based on:

• various didactic domains;
• different techniques of artificial intelligence;
• several countries from different continents;
• number and quality of related publications.

2.4.2. Description of systems

2.4.3. Specification of approaches

2.4.4. Study results and discussion

The framework analysis identifies the following main drawbacks of existing adaptive construction approaches. It should be noted that construction approaches are not sufficiently automated to automatically derive the achievement of objectives from the strategic process.

For the nature view, the different systems do not adopt all pedagogical methods to achieve the different types of process outcomes. In fact, the outcomes attitude and motor skill are not taken into consideration. In addition, the different systems support the tactic and linear processes of learning.

For the description view, most of the systems try to support the learner model, the domain model and the pedagogical model by applying proprietary notation, which can manifest some standardization problems. This notation is not used to specify a process description by respecting the activity level. Indeed, the context coverage is not supported by the learning process description.

For the design view, the uncertain and the evolved context of construction are not supported by the systems. The different on-the-fly approaches do not support a combined construction technique of learning processes to implement the instantiation methods. Most of these approaches are not implemented by applying a MAP-based model of optimization. For that, the educational intentions are not defined as optimization parameters. The construction approaches themselves are not sufficiently guided by considering the different methods and parameters. In fact, the flexible guidance is most used by the expository methods. However, this flexible guidance does not adopt the intention as outcome by considering the learning style and the teaching style.

Finally, the different systems do not apply micro methods by respecting system information to support process adaptation.

Therefore, the comparative study identifies the following limits:

– the adaptive ability of most of these ITS is limited to the content. In fact, the process-oriented adaptation is not supported by the different systems;

– the consideration of the educational process as a strategic and not linear one is limited;

– the flexible guidance of educational processes does not adopt the intention as outcome by considering the learning style and the teaching style;

– the different systems did not respect the evolving process policy.

1. Badaracco, M. and Martínez, L. (2013). A fuzzy linguistic algorithm for adaptive test in intelligent tutoring system based on competences. Expert Systems with Applications, 40(8), 3073–3086.
2. Bayounes, W., Saâdi, I.B., Kinshuk, K., Ben Ghézala, H. (2012). Towards a framework definition for learning process engineering supported by an adaptive learning system. Proc. ICTEE Conference, Amritapuri, India, 366–373.
3. Bayounes, W., Saâdi, I.B., Kinshuk, K., Ben Ghézala, H. (2013). An intentional model for learning process guidance in adaptive learning system. Proc. 22nd IBIMA Conference, Rome, Italy, 1476–1490.
4. Bayounes, W., Saâdi, I.B., Kinshuk, K., Ben Ghézala, H. (2014). An intentional model for pedagogical process guidance in adaptive learning system. Proc. 23rd IBIMA Conference, Valencia, Spain, 1211–1227.
5. Bayounes, W., Saâdi, I.B., Ben Ghézala, H. (2019). Educational processes’ guidance based on evolving context prediction in intelligent tutoring systems. Universal Access in the Information Society, 8(68), 1–24.
6. Bayounes, W., Saâdi, I.B., Ben Ghézala, H. (2020). Definition framework of educational process construction supported by an intelligent tutoring system. Proceedings Multi-Conference OCTA (Organization of Knowledge and Advanced Technologies), Tunis, Tunisia [Online]. Available at: https://multiconference-octa.loria.fr/multiconference-program/.
7. Buche, C., Bossard, C., Querrec, R., Chevaillier, P. (2010). {PEGASE}: A generic and adaptable intelligent system for virtual reality learning environments. IJVR, 9(2), 73–85.
8. Butz, C., Hua, S., Maguire, R.B. (2004). Bits: A Bayesian intelligent tutoring system for computer programming. WCCCE ‘04, 179–186.
9. Chang, M., D’Aniello, G., Gaeta, M., Orciuoli, F., Sampson, D., Simonelli, C. (2020). Building ontology-driven tutoring models for intelligent tutoring systems using data mining. IEEE Access, 8, 48151–481621.
10. Dunn, T.J. and Kennedy, M. (2019). Technology enhanced learning in higher education; Motivations, engagement and academic achievement. Computers & Education, 137, 104–113.
11. Forehand, M. (2012). Bloom’s taxonomy. In Emerging Perspectives on Learning, Teaching, and Technology, Orey, M. (ed.). CreateSpace, Scotts Valley, CA, USA.
12. Garcia-Barrios, V.M., Mödritscher, F., Christian, G. (2004). The past, the present and the future of adaptive e-learning: An approach within the scope of the research project AdeLE. Proc. ICL Conference, Villach, Austria, 1–9.
13. Glass, M. (2001). Processing language input in the CIRCSIM-tutor intelligent tutoring system. Artificial Intelligence in Education, 210–221.
14. Kumar, A. and Ahuja, N.J. (2020). An adaptive framework of learner model using learner characteristics for intelligent tutoring systems. In Intelligent Communication, Control and Devices, Choudhury, S., Mishra, R., Mishra, R.G., Kumar, A. (eds). Springer, Singapore.
15. Merrill, M.D. (1983). Component display theory. Instructional Design Theories and Models: An Overview of their Current Status, Reigeluth, C.M. (ed.). Routledge, New York, USA.
16. Mitrovic, A., Martin, B., Mayo, M. (2000). Using evaluation to shape ITS design: Results and Experiences with SQL-Tutor. Int. J. User Modeling and User-Adapted Interaction, 12(2–3), 243–279.
17. Olesea, C. (2019). The concept of personal learning pathway for intelligent tutoring system GeoMe. Computer Science Journal of Moldova, 27(3), 355–367.
18. Paramythis, A. and Loidl-Reisinger, S (2004). Adaptive learning environments and e-learning standards. Electronic Journal of Elearning, 2(1), 181–194.
19. Patel, A. and Kinshuk, K. (1997). Intelligent tutoring tools in a computer integrated learning environment for introductory numeric disciplines. Innovations in Education & Training International, 34(3), 200–207.
20. Rathore, A.S. and Arjaria, S.K. (2020). Intelligent tutoring system. In Utilizing Educational Data Mining Techniques for Improved Learning: Emerging Research and Opportunities, Bhatt, C., Sajja, P., Liyanage, S. (eds). IGI Global, Hershey, PA, USA.
21. Rolland, C. (1998). A comprehensive view of process engineering. Proc. 10th CAISE Conference, Pisa, Italy, 1–24.
22. Rumelhart, D.E. and Noorman, D.A. (1978). Accretion, tuning, and restructuring: Three modes of learning. In Semantic Factors in Cognition, Cotton, W. and Klatzky, R. (eds). Erlbaum, Hillsdale, NJ, USA.
23. Tzanova, S. (2020). Revolution by evolution: How intelligent tutoring systems are changing education. In Revolutionizing Education in the Age of AI and Machine Learning, Habib, M. (ed.). IGI Global, Hershey, PA, USA.