Method and system for predicting oncomelania density based on ReliefF-SVM

Abstract

The invention relates to an oncomelania density prediction method and system based on a ReliefF-SVM. The oncomelania density prediction method comprises the steps of obtaining oncomelania situation data; the oncomelania condition data comprises the position of an oncomelania breeding place and the average density of live oncomelania; acquiring geographical environment characteristic data of the position of the oncomelania breeding place; the geographical environment characteristic data comprises meteorological data, soil texture data, soil type data, water system data, vegetation type data and vegetation coverage data; constructing a sample set based on the average density of the live snails and the geographical environment feature data; on the basis of a ReliefF algorithm, selecting each feature in the geographical environment feature data in the sample set; training a support vector machine by using the sample set after feature selection to obtain an oncomelania density prediction model; and predicting the oncomelania density by using the oncomelania density prediction model. The oncomelania density prediction is carried out based on the ReliefF-SVM combination model, and the prediction accuracy and the prediction efficiency are effectively improved.

Description

technical field [0001] The invention relates to the technical field of schistosomiasis control and prevention, in particular to a method and system for predicting oncomelania density based on a ReliefF-SVM combination model. Background technique [0002] Schistosomiasis is an infectious disease on the surface of the body. The only intermediate host of schistosomiasis is the snail, and the miracidia in the snail will develop into cercariae with infectious ability. The distribution and living environment of Oncomelania snails are the most important factors affecting schistosomiasis. Therefore, accurate identification of Oncomelania breeding grounds plays a key role in grasping the distribution of schistosomiasis and preventing and controlling schistosomiasis. [0003] There are many types of infectious disease prediction methods, and the classification methods are different. In recent years, neural network models have been applied to the prediction of infectious diseases. Ar...

AI Technical Summary

Problems solved by technology

[0004] These methods show that support vector machines can be better applied in the field of infectious disease monitoring, and most of the monitoring research on oncomelania breeding sites only use support vector machine classification to qualitatively determine whether there are living areas of oncomelania, but do not quantitatively predict the occurrence of oncomelania. Density, so the present invention proposes a method and system for predicting oncomelania density based on the ReliefF-SVM combination model to realize quantitative research on oncomelania density

Images