Learning Algorithms

Table

• Profile Initialization
• Lambda and Weight updater
• Profiles update

We initiate a pool of models that will pass individually through the learning algorithms to improve their's efficiency using Sobrie's metaheuristic:

• ProfileInitializer initializes the profiles
• WeightUpdater update the weights of the criteria
• ProfileUpdater update the profiles given the fixed weights

Then, we regroup the models, sort them by accuracy and reinitialize the last half of the models.

After n iteration or an accuracy of 70%?? for each model, the algorithm is stopped.