Title :
Combining probabilistic neural networks and decision trees for maximally accurate and efficient accident prediction
Author :
Tambouratzis, Tatiana ; Souliou, Dora ; Chalikias, Miltiadis ; Gregoriades, Andreas
Author_Institution :
Dept. of Ind. Manage. & Technol., Univ. of Piraeus, Piraeus, Greece
Abstract :
The extent to which accident severity can be predicted from accident-related data collected at a variety of locations is investigated. The 2005 accident dataset brought together by the Republic of Cyprus Police is employed; this dataset comprises 1407 records of 43 continuous and categorical input parameters and a single categorical output parameter representing accident severity. No transformation of the database has been opted for, either by extracting the parameters that are significant for the prediction task or by modifying the records in any way (e.g. via record selection or transformation). Aiming at maximally accurate and efficient prediction, a combination of probabilistic neural networks (PNN´s) and decision trees (DT´s) is implemented: the simple training and direct operation of the PNN is complemented by the hierarchical, exhaustive and recursive construction of the DT. By training pairs of PNN´s on data from the partitions derived from the minimal necessary number of top DT nodes, both efficiency and accident prediction accuracy are maximized.
Keywords :
decision trees; neural nets; road accidents; traffic engineering computing; Republic of Cyprus Police; accident prediction; accident severity; decision trees; probabilistic neural networks; Accidents; Boolean functions; Data structures; Estimation; Predictive models;
Conference_Titel :
Neural Networks (IJCNN), The 2010 International Joint Conference on
Conference_Location :
Barcelona
Print_ISBN :
978-1-4244-6916-1
DOI :
10.1109/IJCNN.2010.5596610