Polish Version
Kinga Kądziołka
Summary
This article attempts to model the daily number of road accidents in Poland using data mining methods such as random forests and artificial neural networks. There was compared the network, which takes into account all the analyzed explanatory variables to the network with reduced the number of input variables. The network with reduced set of input variables was characterized by a slightly lower average absolute percentage error on the test set than the network, which includes all explanatory variables. There was also identified clusters of poviats characterized by high risk of road accidents.
Spatio – temporal analysis of road accidents in Poland
Article
In order to view this page you need Flash Player 9+ support!
