Multi-model fusion based on MATLAB to identify the water quality between the wet season and the drought period in Changyi water source area

Water quality is an important index to reflect hydro-environment of the drainage basin. Aiming the problem of lower accuracy in traditional water quality recognition methods, Changyi water source area was taken as an example, and ten kinds of hydrochemical component indexes were selected as the basic indicators for water quality identification. Based on the SPSS factor analysis modeling, the water quality indexes closely related to first principal factor, which were determined among the twenty groups of training samples of wet season and the drought period. Then, influencing weights of main controlling factors found closely related to the first principal factor which were confirmed by the combination of variance analysis and coefficient of variation methods, based on D-S improvement theory of conflicting evidence. Thereby, according to distance discriminant analysis algorithm, the obtained water quality identification indexes and their weights were input into the MATLAB platform as the sample data, results in the Mahalanobis distance matrices between training samples of the wet and drought seasons, which were eventually determined by using sqrt and mahal functions. Further, eight unknown test samples were entertained. Examples prove that when the identification indexes were selected appropriately, the multi-model fusion based on MATLAB was used to identify the water quality between both time periods in Changyi water source area. It effectively eliminate the mutual influences among the discriminant indexes and improve the discriminant rate, which further provides a theoretical basis for rational development, utilization, and protection of water resource as well as evaluation of the utility value of water system.