Portal de Periódicos da CAPES

Judgement of Palatable Water Using Fuzzy Reasoning.

1996; Japan Society on Water Environment; Volume: 19; Issue: 3 Linguagem: Inglês

10.2965/jswe.19.209

1881-3690

Chihiro IWANAGA, Ken Sasaki,

Fuzzy Logic and Control Systems

We propose applying the fuzzy reasoning with four input variables (organic matter, total hardness, iron, bicarbonate) to the evaluations of palatable water.Firstry, the membership functions of them were prepared from the standard on palatable water from Ministry of Welfare in Japan. After that, a learning rule based on a desent method is applied to the consequent part of each fuzzy rule. The usefulness of generated fuzzy rules for test data (N=35, other district) was examined, so the correlation coefficient was given as 0.7829. Furthermore, we generated new rules iteratively in the fuzzy partitioned input regions by tuning. The correlation between sensory test and evaluation by fuzzy rules with the minimum inference error was improved on and the correlation coefficient was given as 0.8621.These results suggest it is possible to apply fuzzy if-then rules with non-fuzzy singletons in the consequent part and four input variables to the evaluations of palatable water.