Anna Chmielińska-Bernacka, Robert Sidełko
Politechnika Koszalińska

In article the results of neural network (NN) application connected with prediction of amount of municipal waste was presented. To neural network training about one hundred cases was used. One case mean household where a individual questionnaire enclosed fourteen questions was filed up. Finally only nine features treated as independent variables was qualified. Neural network verification depended on comparing real and predicted value for two cases removed during NN teaching process.

Słowa kluczowe
sieci neuronowe; odpadóy bytowogospodarcze

Pełny text / Full text