The Strength of Weak Learnability
Machine Learning
Data mining: practical machine learning tools and techniques with Java implementations
Data mining: practical machine learning tools and techniques with Java implementations
Artificial Neural Networks in Hydrology
Artificial Neural Networks in Hydrology
| Hi-index | 0.00 |
Machine learning and soft computing techniques in application to water resources allows for more accurate prediction of water flows, for example in flood situations. In particular, ANNs in hydrological modelling demonstrated their high applicability. There are, however, certain problems of their practical use: ANNs are typically trained on the whole data set and being accurate on average may miss the extremes; their internal structure cannot be interpreted by practitioners. A solution is seen in the use mixtures of models that can be easier interpreted. The paper discusses the practical applications of hierarchical model structures (M5 model trees) and their comparison to ANNs. It is demonstrated that model trees, being almost as accurate as ANNs, may better fit the practical needs of hydrologists in flood related problems.