Neural networks and genetic algorithms for hierarchical multi-label classification problems

In conventional classification problems, each instance of a dataset is associated with just one among two or more classes. However, there are more complex classification problems where instances can be simultaneously classified into classes belonging to two or more paths of a hierarchy. Such a hierarchy can be structured as a tree or a directed acyclic graph. These problems are known in the machine learning literature as hierarchical multi-label classification (HMC) problems. In this Thesis, two methods for hierarchical multi-label classification are proposed and investigated. The first one associates a Multi-Layer Perceptron (MLP) to each hierarchical level, being each MLP responsible for the predictions in its associated level. The method is called HMC-LMLP. The second method induces hierarchical multi-label classification rules using a Genetic Algorithm. The method is called HMC-GA. Experiments using hierarchies structured as trees showed that HMC-LMLP obtained classification performances superior to the state-of-the-art method in the literature, and superior or competitive performances when using graph-structured hierarchies. The HMC-GA method obtained competitive results with other methods of the literature in both tree and graph-structured hierarchies, being able of inducing, in many cases, smaller and in less quantity rules