Skip to main content

Table 4 Identification of most influential predictor variables for the best 'cross-location' and 'cross-year' SVM models.

From: Machine learning techniques in disease forecasting: a case study on rice blast prediction

Weather variable(s) excluded

Best SVM 'cross-year' model Correlation coefficient (r)

Best SVM 'cross-location' model Correlation coefficient (r)

None excluded. All 6 variables (Tmax, Tmin, RHmax, RHmin, Rainfall, RD/week) included

0.979

0.982

Rainfall

0.789

0.949

Rainy days/week

0.974

0.966

Relative Humidity (minimum)

0.866

0.982

Relative Humidity (maximum)

0.953

0.984

Temperature (minimum)

0.968

0.983

Temperature (maximum)

0.973

0.984