Risk factors database
Risk factors such as weather parameters from different sources includes the monthly precipitation(mm), sea level pressure (millibar),minimum temperature (°C), maximum temperature(°C), wind speed (m/s), vapour pressure (millibar), soil moisture(%) , perceptible water(mm), potential evaporation transpiration (mm), cloud cover(okta) etc., extracted from various databases such as National Centre for environmental prediction (NCEP), Indian Meteorological Department(IMD),National Innovations Climate Resilient Agriculture (NICRA) and other sources. The remote sensing variables like Normalised Difference Vegetative Index (NDVI) and Land Surface Temperature(LST) were extracted from MODIS/LANDSAT/LISS III or IV satellite images. The livestock population and densities were extracted from Livestock census 2012.
Artificial Intelligence Algorithm
Two regression models viz. Generalized linear models(GLM) and Generalized additive models (GAM) and six Machine learning methods viz. Random Forest (RF), Boosted Regression Tree (BRT), Artificial Neural Network (ANN), Multiple Adaptive Regression Spline(MARS), Flexible Discriminant Analysis (FDA) and Classification Tree Analysis (CTA). Different modelling methods return different type of `model object’, all these model objects could be used to with the predict function to make predictions for any combinations of values of independent variables. Response plots were created to explore and understand model predictions. To assesses the models for reasonable predictions and spatial patterns, different measures are used to assess how good the model fits the data.
Difference between NADRS and NADRES
NADRS is the acronym for National Animal Disease Reporting System, which is an online
animal disease reporting system developed by Dept of Animal Husbandry Dairying and
Fisheries, Ministry of Agriculture, Govt of India. In this system, 7500 block level veterinary
hospitals across India are electronically networked for real time disease reporting. The
disease data entered in the system are analysed to understand the livestock disease status
across India.
The National Animal Disease Referral Expert System (NADRES) developed by ICAR-
National Institute of Veterinary Epidemiology and Disease Informatics (ICAR-NIVEDI) uses
the livestock disease data so generated by the DADF & its own 31 AICRP centers for
forewarning of 13 predetermined (most reported) livestock diseases two months in
advance. Briefly, the livestock disease data aggregated at the district level is integrated with
24 other climatic [land surface temperature (LST), precipitation, wind velocity, humidity
etc.,] and non climatic parameters [such as district level livestock population, density,
Normalized Differential Vegetation Index (NDVI), soil moisture etc], which are
extracted/collected from various sources including remote sensed images, in the server
located at ICAR-NIVEDI. Following the data integration, sophisticated statistical models are
run in R environment using R programming language and districts are classified into six
different risk categories for each of the thirteen diseases. The forewarning outputs are
generally depicted both in tables and GIS maps. The so generated forewarning information
is circulated to both DADF, GOI and the state animal husbandry department two months in
advance.