Logistic coefficients and correlation

59 aspect and slope have contributed more cells that being assumed for deforestation occurrence. Slope has contributed 83.8 of value 1 as indicated tends to be deforested, and aspect contributed 43.4.

4.3.3. Logistic coefficients and correlation

These are the values for the logistic regression equation for predicting the dependent variable from the independent variable. They are in log-odds units. Similar to Ordinary Least Square OLS regression, the prediction equation is logp1-p = -0.238 + LogR_Riv-0.232 + LogR_Road1.130 + LogR_SL0.348 + LogR_CP0.354 + LogR_COM0.082 + LogR_Slp-0.150 These estimates tell us about the relationship between the independent variables and the dependent variable, where the dependent variable is on the logit scale. These estimates tell the amount of increase or decrease, if the sign of the coefficient is negative in the predicted log odds of deforestation = 1 that would be predicted by a 1 unit increase or decrease in the predictor, holding all other predictors constant. This testing can do this by hand by exponentiating the coefficient, or by looking at the right-most column in the Variables in the Equation and Wald Test Table 4.8 labeled Exp β. Odd ratio, in this study is showing that distance less than 1,000 km tends to deforested occurrence 3.097 times than distance greater or equal 1 km from existing road. The smallest possibility of deforestation occurrence was contributed by predictor distance 1 km from river, and almost has no effect to deforested occurrence. In another word, odds ratio close to 1.0 indicate that unit changes in that independent variable do not affect the dependent variable. To test each coefficient of model beside Exp β, there is Confidence interval on the odds ratio itself Table 4.8. When the 95 confidence interval around the odds ratio includes the value of 1.0, indicating that a change in value of the independent variable is not associated in change in the odds of the dependent variable assuming a given value, then that variable is not considered a useful predictor in the logistic model. Classplot in Figure 4.21 is an alternative way for assessing correct and incorrect prediction under logistic regression. The X axis is the predictor for probability from 0 to 1. 0 of the dependent variables being classified “changed forest”. The Y axis is frequency: the number of cases classified. Inside the plot are column of observed deforestation’s and unchanged’s which it here codes as 1’s and 0’s, with 200 cases per symbol. Examining this plot will tell such things as how well the model classifies difficult cases. In this case, it also shows nearly all cases are coded as being in the 0 unchanged forest and non forest group, and only few in the 1 deforestation group. 60 Predicted Probability is of Membership for 1 The cut Value is 0.50 Symbols: 0 – 0 1 – 1 Each Symbol Represents 200 cases Figure 4.17. Classification plot ClassPlot, another very useful piece of information for assessing goodness of fit for the model. 61 4.4. Validation and Accuracy Assessment 4.4.1. Validation