Predicting the Density of Algae Communities using Local Regression Trees
Luís Torgo
1999
Abstract
This paper describes the application of local regression trees to an environmental regression task. This task was part of the 3rd Internation ERUDIT Competition. The technique described in this paper was announced as one of the three runner-up methods by the jury of the competition. We briefly describe RT, a system that is able to generate local regression trees. We emphasise the multi-strategy features of RT that we claim as being one of the causes for the obtained performance. We described the pre-processing steps that were taken in order to apply RT to the competition data, highlighting the weighed schema that was used to combine the predictions of the best RT variants. We conclude by reinforcing the idea that the combination of features of different data analysis techniques can be useful to obtain higher predictive accuracy