3.4.6 Factors influencing flea abundance
A best final fitted model obtained from the GLM (Poisson family) was statistically significant to explain the relationship of predictor variables to the abundance of flea χ² (15) = 106.63, p< .001, model coefficients in (table 3) , predictors chart in fig. 4 (4B-4F) showing the probability of fleas in different factors.
Figure 4 . (A) Scatter plot for Spearman’s rank correlation test of rodent’s weight and flea abundance. (B) is the GLM prediction curve of influence of rodent’s weight on abundance of fleas. (C-F) are the GLM categorical predictor variables with their influence to abundance of fleas.