The invention discloses a
water quality soft measuring prediction method of
permanganate index, which relates to the field of environment monitoring. The method comprises the steps of obtaining the
data value of a plurality of
water quality monitoring indexes in
water environment to be measured, wherein the
data value obtained by each
water quality monitoring index is more than one, and the plurality of water
quality monitoring indexes include at least one
permanganate index; selecting a plurality of water
quality monitoring indexes from various water
quality monitoring indexes excluding the
permanganate index, and forming a data sample set with the
data value of the selected water quality monitoring index, and dividing the data sample set into a
training set and a testing set, conductingtraining on the
training set by means of
support vector machine algorithm, and obtaining a
permanganate index soft measuring model; conducting test on the
permanganate index soft measuring model by means of the testing set, and obtaining the testing result; repetitively executing model establishing and
model testing until the obtained test result accords with the preset condition, and using the
permanganate index soft measuring model according with the preset condition as the final soft measuring model. The invention is advantageous in that the prediction effect of the model is accurate, and the water quality soft measuring prediction cost of the permanganate index can be reduced.