Development of a Classification Model for the Assessment of Maize Plant Yield in Nigeria

Abstract

The identification of diseases in plants is usually achieved through the interpretations made of the visual symptoms by an agricultural expert. However, in situations where such experts are unavailable or the farmer’s knowledge is insufficient, other methods for field-based assessments are a critical need. The need for a classification model to carry out field-based assessment of the yield of maize plant based on severity of symptoms informed this research. The study proposes a fuzzy-based model with triangular membership functions for the fuzzification of risk factors which were identified by experts of maize plant yield. Thirty-two rules were inferred using IF-THEN statements which adopted the values of the associated risk factors as antecedent and the yield of maize plant as the consequent part. The fuzzy logic model was simulated using five risk factors as input variables and the plant’s yield as the output variable. The results showed that associated risk factors include the presence of black mould growth; blights on leaves, rots on cobs, infected husks and black kernels, and seed decay have noticeable influence on the yield of maize plant. The study concluded that the lesser the presence of such risk factors, the higher the yield of the maize plant.