The artificial neural network model predicts the number of dengue cases based on weather factors in Vung Tau City of Ba Ria–Vung Tau province: a correlational study

  • Quang Minh Lam
  • Hao Thi Tu Vu
  • An Dao Thien Nguyen
  • Dang Ngoc Tran

Tóm tắt

This study was conducted to build an early forecast model for the dengue epidemic using an artificial neural network (ANN) in Vung Tau City in Ba Ria - Vung Tau province. Weather factors (temperature, precipitation, humidity, wind speed) were all correlated with the number of dengue cases (p<0.05) in the 1/2010 - 12/2020 period. The ANN is a simulated information processing model based on the activity of the organism's nervous system, consisting of a large number of neurons mounted to process information. Of all the models we built, the ANN 10-7-1 model gave the best results with a total squared error (SSE) and relative error (RE) during network training of SSE=25.23, respectively; RE=0,58; correlation coefficient R2 =29.1%. The model showed that the number of dengue cases have marked tendencies to increase to increase or to decrease, also the forecast data was almost accurate compared to the actual number of dengue cases. The ANN can be applied to predict future dengue outbreaks in Vung Tau City. To increase the accuracy of the ANN model, it needs to be combined with other factors such as housing index, mosquito density, population,... 

Full text link

điểm /   đánh giá
Phát hành ngày
Chuyên mục
Bài viết