Laserfiche WebLink
i~ <br />'~ <br />proportions of misclassified site locations (false negatives). For <br />example, the Linear Discriminant Function (LDF) based on slope, <br />' classified 86 percent of site locations and 82 percent of nonsite <br />locations; overall 84 percent of cases were classified correctly. <br />The LDF based on slope and distance to water correctly classified <br />' 91~ of the site sample but misclassified nearly 50$ of the <br />nonsites. These results are summarized in Table 10. <br />' The prediction equation for slope, vegetation and land form <br />is: <br />' [(0.57*Slope) + (0.28*Vegetation) + (0.83*Landform)]-3.04 <br />where slope, vegetation and land form are standardized scores using <br />' the codes listed for nominal variables. <br />The prediction equation for slope and distance to water is: <br />[(0.50*Slope) + (0.009*Water)}-3.26 <br />where slope and water are standardized scores for those variables. <br />For both equations discriminant scores less than 1 indicate a <br />' high probability of site presence while scores greater than 1 <br />indicate a high probability of site absence. <br /> <br />1 <br /> <br /> <br /> <br /> <br /> <br />86 <br /> <br />