Sciweavers

Share
ICDM
2009
IEEE

Induction of Mean Output Prediction Trees from Continuous Temporal Meteorological Data

10 years 10 months ago
Induction of Mean Output Prediction Trees from Continuous Temporal Meteorological Data
: In this paper, we present a novel method for fast data-driven construction of regression trees from temporal datasets including continuous data streams. The proposed Mean Output Prediction Tree (MOPT) algorithm transforms continuous temporal data into two statistical moments according to a user-specified time resolution and builds a regression tree for estimating the prediction interval of the output (dependent) variable. Results on two benchmark data sets show that the MOPT algorithm produces more accurate and easily interpretable prediction models than other state-of-the-art regression tree methods. Key words: temporal prediction; inductive learning; time resolution; regression trees; split criteria; multivariate statistics; multivariate time series
Dima Alberg, Mark Last, Roni Neuman, Avi Sharon
Added 23 May 2010
Updated 23 May 2010
Type Conference
Year 2009
Where ICDM
Authors Dima Alberg, Mark Last, Roni Neuman, Avi Sharon
Comments (0)
books