The main objective of WP4 is to develop the student models, thus realising the LA. This is divided into the following sub-objectives:
- Study and select tools and techniques for the realisation of the student models, based on the requirements of WP3 and the outputs of WP2
- Define the architecture and APIs with respect to scalability and interfaciality to the portfolio system
- Implement and performance test the student model architecture
DESCRIPTION OF WORK AND ROLE OF PARTNERS
Description of Work: This WP includes the same Partners as WP 3. UR is leading this WP. Both UR and UM have specialised knowledge in the development of student models. Their roles are mainly in T4.1 and T4.3. The API design will be worked out by the SMEs, especially NetRom. SME JW functions as a linking pin to WP 5 in which it has the lead.
Task 4.1 Selection of Tools and Techniques for Student Model: Based on the input from WP2 on variables of the knowledge domain and other technical considerations such as scalability, this task will select the educational mining tools and techniques that will be deployed to learn, update and store the student models.
Already, several Bayesian Belief Networks (such as multi entity belief networks (MEBN), relational belief networks and rich probability models) are able to make use of this type of relational information. Thus, such approaches, especially MEBNs, will be a likely candidate for realising the student model. The student model will thus realise the cognitive model, i.e. representing the current status of the student in the domain.
In addition, context awareness is a major factor in work place-based education as the influence of current circumstances, such as current cases to be worked on, changes in supervision, etc. greatly impact the student performance and thus the student model. Therefore, context awareness approaches will be studied and the most suitable one selected, taking into account the access to data (as identified in WP3) necessary for context awareness.
Finally, to generate feedback to both student and instructor/teacher, tools and techniques for (training) content (re-) structuring can be used and adapted towards the work place-based training setting to allow for suggestions to be made as to on what to focus in additional training.
Task 4.2 Architecture & API Design: In order to enhance the reusability of the project results and to allow independent development within Work Packages within the project, it is wise to develop an architecture that separates student models from the current assessment data in the portfolio system or for that matter any other related system. Moreover, it is expected that information only flows from the portfolio system application to this module and not in the opposite direction. Given the expected computational resources needed, this module might run on a different platform or machine than the portfolio system application itself.
The task will therefore establish the architecture for the Student Model Module including a development and deployment platform, software tools as well as APIs to allow for re-usability. The specification will include:
- Design of communication between portfolio system and the module e.g. push or pull information exchange
- Interface to the portfolio system database
- API (programming interface) for generating output data for feedback and visualisations in Work Package 4
- Mock-up for this API in order to allow Work Package 4 to start developing.
- Techniques and algorithms for student model generation and online update of models
- Efficient storage of student models such that access to and update of the models can appear as fast as possible.
- Context Awareness sub-system to efficiently update the student model with the latest information.
- Suggestion generation based on current status using existing techniques and technologies for structuring training content, adapted to work place-based training
- Test plan: what aspects of the system should be tested
Task 4.3 Implementation of the Student Model: This task will focus on the implementation of the student module as per the specification of Task 4.2. This will also include the initiation/ population of the starting model. The task will also realise initial performance tests to ensure scalability as well as conformance with the requirements.