Objective To apply several comprehensive methods to the evaluation of hygienic conditions of swimming places, compare the reasonability of these methods, and test their consistency.
Conclusion Two kinds of data were obtained from the archives of the health supervision and water quality monitoring of thirty swimming places in Changning District, Shanghai in 2009. A quantitative and graded management method was used for health supervision data. Models based on four methods, i.e. qualification rate, comprehensive index, fuzzy mathematics, and artificial neural network, were built to rank water quality levels. Results of five comprehensive models were tested for consistency.
Results Compared with the other evaluation methods, the advantage of artificial neural network method lay in methodology. The evaluation model based on artificial neural network method showed that the average level of water quality was 2.27 and that the water quality level distribution of swimming places was spindle shaped. Moreover, the proportion of swimming places at excellent and general water quality levels in gymnasiums (100.0%) was higher than those in community health clubs (68.2%). These results were basically consistent with health supervision data over the past years. The average level of water quality indicated by models based on comprehensive index method, qualification rate method, and fuzzy mathematics method was 3.23, 2.87, and 2.20, respectively. The average level of swimming places evaluated by quantitative and graded management method was 2.27. The coefficients of Kappa ranged from-0.134 to 0.173 between the results of the five comprehensive evaluation models.
Conclusion There is a low consistency between the results of the five comprehensive evaluation methods. The model based on artificial neural network method is relatively reasonable.
DownLoad: