Self-Organizing Maps for Mixed Feature-Type Symbolic Data
Résumé
The self-organizing map is a kind of artificial neural network used to map high dimensional data into a low dimensional space. This paper presents a self-organizing map to do unsupervised clustering for mixed feature-type symbolic data while preserving the topology of the data. A preprocessing technique prior to clustering is needed in order to homogenize the data. Every mixed feature-type vector is transformed into a vector of histograms. The resulting data set is used to train the self-organizing map using the batch algorithm. Similar input vectors will be allocated to the same neuron or to a neighbor neuron on the map. The performance of this approach is then illustrated and discussed while applied to real interval and mixed feature-type symbolic data sets.