Annealing Based Approach to Optimize Classification Systems

Classification systems optimization is often performed with population based genetic algorithms. These methods are known for their efficacy on solving these problems, but are associated to a high computational burden with classification systems. This paper evaluates an annealing based approach to optimize a classification system, and compares results obtained with a multi-objective genetic algorithm in the same problem. Experiments conducted with isolated handwritten digits demonstrate the effectiveness of the annealing based approach, which encourages further research in this direction.