Influences Of Clustering Modifications On The Performatnce Of The Genetic Algorithm Driven Clustering Algorithm

One way to look at basic modeling approaches is to split them up into mechanistic and data based models. A few years ago we developed our own data based model approach [1], called Genetic Algorithm driven Clustering (GAdC). The proposed methodology relies on semisupervised clustering with a generative floating-point genetic algorithm and local learning. In this contribution we investigate the influence of clustering modification on the performance of the prediction of a real world application [2]. The task is the prediction of alga frequency distributions on the basis of the measured concentrations of the chemical substances, the global information concerning the season when the sample was taken, the river size and its flow velocity.