Аннотация:As the water quality of Chao Lake becomes diverse and eutrophied, predicting and evaluating water quality become more and more important. But prediction and evaluation are complex problems. In this paper, we come up with an improved decision tree learning method making water quality prediction easier and forecast more accurate. The classification standards are based on the evaluation mechanisms provided by the Hong Kong Environment Protection Department. We released an online web forecast system to apply to the classification and prediction of Chao Lake. Experimental results show that the improved method is better than artificial neural network or genetic algorithm with higher recognition rate and forecast accuracy and strong practical value.
