Multiple Linear Regression Fitting six variables with eight data points is not recommended as it artificially inflates the coefficient of determination. The full model Y~X1+X2+ X3 + X4 + X5+X6 shows an ; however, nothing is significantly contributing as some of the variables are confounded. Moreover, I can make by including the interaction between X3 and X6; however, this is an exact solution with no measure of error - eight equations and eight unknowns. Consider the model Y~ X1 + X4 + X5 that is, Y ̂ =11.2+0.35X1-0.24X4-4.42X5;R 2 =0.7311 This model explains 73.11% of the variance in Turbidity; that is, approximately 73% of the average turbidity is explained by the Developed, Wetland and Bare land with all variables found to be significant at the 5% level.
PCA revisited PC1PC2PC3PC4 DevelopedX10.503-0.861 WetlandX40.8650.5 Bare landX5-0.1650.982 Turbidity: All stations (turb <24)Y0.9860.163