Gradual model generator for single-pass clustering

We present an algorithm for generating a mixture model from data set by performing a single pass over the data. The method is applicable when the entire data is not available at the same time in the main memory. We use Gaussian mixture model but the algorithm can be adapted to other types of models, too. We also outline a post processing method, which can iteratively reduce the size of the model obtained by the single-pass algorithm. This results in a model with fewer components, but with approximately the same representation accuracy than the result of the original model from the single-pass algorithm.