Soft data analysis for evaluating regional environment

There is a plenty of statistical data related to environmental issues nowadays, which is called hard data in this paper. These numerical data, however, cannot cover all information required in the studies of environmental problems. Sometimes opinions of residents are much more informative than the hard data. The subjective data obtained by a questionnaire survey is called soft data in this paper. This paper tries to identify the relation between hard data and soft data with the fuzzy modeling techniques. After introducing a questionnaire survey we have carried out, the paper presents a classification of districts based on hard data. Then, a rule-based environmental evaluation model is developed in each cluster. This process includes a new fuzzy modeling technique to deal with soft data. Especially, a selection method of data points as well as indicators or attributes is proposed by consulting the data mining technique