Prediction of Echinococcosis Transmission in Qinghai Province Based on Machine Learning

This article collects data on echinococcosis infection in the population, dogs, livestock, and rodents of Qinghai Province through literature review. Four traditional machine learning methods are used to predict the positive rates of the population, dogs, livestock, and rodents in the next three years based on the collected disease rate data from the past six years (2016–2021). In addition, to expand the dataset of the affected population in Qinghai Province, the annual population disease rate and total number of people in Qinghai Province are analyzed to obtain the total number of people affected each year. The total number of people affected each year is dispersed to each day of the year by using Normal distribution to expand the dataset, and an improved LSTM model is used to fit the expanded dataset. After conducting experimental simulations on the prevalence data, it reveals that the prevalence of echinococcosis in the population, dogs, livestock, and rodents in Qinghai Province has been decreasing year by year since 2022. After conducting experiments on the expanded dataset of affected populations, it shows that the fitting results of the improved LSTM model were 8.7% lower than those of the original LSTM model in terms of MSE. Using the improved LSTM model to predict the daily prevalence of echinococcosis in subsequent years, it reveals that the annual prevalence of echinococcosis in the region is below 0.06 %, indicating that the overall trend of echinococcosis in Qinghai Province is controllable, and the measures taken by Qinghai Province to prevent and treat echinococcosis are effective.