We will concern this part to the initialization of the model of the learner in the AHES, begin by describing the process of initialization of the model that we will propose by combining the Bayesian networks as we presented in the preceding chapters and the stereotypes in order to achieve a complete initialization in all aspects are understood.
The Process of Initializing the Learner Model
The initialization of the learner model represents the process of gathering information about the learner and transferring this information to the model. This process of initialization represents a major problem for adaptive systems.
In this section, we will present our process of collecting data about the learner. According to Self (1991), the learner model can be initialized in three ways, using explicit questions, initial tests, or the method of stereotypes.
Figure 1 shows our proposed process for collecting learner data. We propose in this chapter the use of two methods simultaneously to gather all the information related to the learner, which will give us a more complete learner model, and this will be reflected on the adaptation of the system to the needs of the learner in a more precise way.
Figure 1.
The process of initializing the learner model
To achieve this goal, we will use two methods for data collection: explicit questions, and initial tests.
In order to gather specific information about the learner that makes up the learner's profile, such as personal information, goals, cognitive skills, preferences ... we will base this process on explicit questions, which will guide us in assigning each learner to a well-defined stereotype.
And for independent learner information, which reflects its level of knowledge and skills for each module in the system, and which represent the key element for system adaptation. We base this process on initial tests, using Bayesian networks as a formalism to properly represent and manage this information in a probabilistic way.