Abstract

Secondary data on proximate composition of “fufu” flour taken from forty three cassava mosaic disease (CMD) resistant varieties were used for this work. Agglomerative hierarchical cluster analysis was performed on the squared Euclidean distance matrix. The distance coefficients generated between the forty three CMD resistant varieties ranged from 0.000 to 89.120. Six (6) distinct groups were identified at 0.97 coefficients. A dendrogram of the data indicated that cases with low distances are close together with a line linking them. It was observed that the line was a short distance from the left of the dendrogram indicating that they were agglomerated in a cluster at a low distance coefficient. This indicated likeness. The implication of this to the farmer and indeed to the nutritionist is that a variety can be selected from each of the 6 cluster groups for cultivation with the objective of achieving the same nutritional differences in terms of proximate composition without having to examine all the forty three varieties.

Key words: Cluster, agglomerative hierarchical cluster, squared euclidean distance, cassava “fufu”, CMD-resistant varieties, clustering algorithms.