(Tartu Ülikool, 2014-06-18) Cuevas Urosa, Miguel; Pärna, Kalev, juhendaja; Tartu Ülikool. Matemaatika-informaatikateaduskond; Tartu Ülikool. Matemaatilise statistika instituut
The aim of this work is to study the Nonnegative Least Squares Optimization, to investigate if it is possible to reduce the number of model points in a dataset to save time. We will start with a huge dataset from an insurance company, we are going to optimize this dataset and reduce the number of model point without losing significant accuracy. We do this with the Nonnegative Least Squares (NNLS) method. In this thesis, NNLS will be described briefly, results and conclusions from the NNLS optimization are shown and discussed.