模糊线性回归模型在河流水体总氮浓度预测中的应用

Application of fuzzy linear regression model in predicting river water total nitrogen concentration

以湘江熬洲断面为例,将该断面水体中总氮浓度及其有关影响因子用三角模糊数来表征。同时,结合己有的模糊线性回归模型成果,构造了带有三角模糊参数的水体中总氮浓度模糊线性回归预测模型。并应用所建模型预测该断面水体中2002—2005年总氮浓度,所得的预测值与已有的实测值之间的相对误差均小于20%,完全满足实际应用对误差的要求,预测合格率为100%,说明这种预测模型在预测河流水体总氮浓度变化中有一定的实用性,为今后开展河流水体中污染物浓度预测提供了新途径。

Taking the Aozhou section of Xiangjiang River as a case,the total nitrogen concentration in water body and related affecting factors were characterized with triangular fuzzy number,and the triangular fuzzy parameters were introduced into the existing fuzzy linear regression model to predict the total nitrogen concentration in water body.The relative errors between the predicted and measured values of total nitrogen concentration in the section in 2002-2005 were less than 20%,with a qualified prediction rate being 100%,which suggested that the modified fuzzy linear regression model had definite practicability in predicting the total nitrogen concentration in river water.