Geographical Variation, Path Diagram, and Regression Tree of the Incidence and Severity of Potato Late Blight (Phytophthora infestans): The Observed Pattern in the Major Growing Areas in Benguet Province

There are various multivariate statistical tools for plant disease epidemics to associate multiple epidemiological factors represented by the disease triangle and tetrahedron. However, its applications are limited in Asian countries, let alone in the locality. This descriptive causal–comparative [farm] survey research, therefore, utilized Cluster Analysis, Path Analysis, and Classification and Regression Tree (C&RT). These multivariate tools were used to assess the geographical variations, account the explained variance/ effect size, and model the causal relationship of multiple variables on host variables, environmental factors, and cultural management practices of farmers to visually assessed incidence and severity of late blight of potato farms in major growing areas in Benguet observed in May 2021. There are three (3) clusters of the observed farms that emerged, of which cluster 3 has consistently had the lowest incidence and severity of potato late blight observed. Therefore, the host characteristics and cultural management observed in cluster 3 are worthy of consideration in potato production. Both the best/tuned regression trees and the combined recursive path diagram showed host variables to have contributed a significantly high proportion of effect/ explained variance and emerged most of the significant predictors for the observed incidence and severity of potato late blight.