In this study, the polynomial fitting curves are used to generate the polynomial function of each land zoning category and translate them into probability functions for the IE formulas (6) and (7) in Section 2.7 to compute their values of conditional entropy and the IG. Table 4 lists the polynomial probability function of the gross dataset and five land zoning categories, where their R2 values reflect the high goodness-of-fit of the polynomial function to the sample distribution. R2 can take any value between 0 and 1, standing for the goodness-of-fit from low to high.
Table 4
Polynomial probability functions
| Category | Polynomial probability functions | R2 |
|---|---|---|
| Gross dataset |  | 0.9992 |
| High-density residential and commercial |  | 0.9995 |
| Low- to medium-density residential |  | 0.9998 |
| Industrial |  | 0.9993 |
| Urban green land |  | 0.9993 |
| Urban road |  | 0.9996 |
| Category | Polynomial probability functions | R2 |
|---|---|---|
| Gross dataset | | 0.9992 |
| High-density residential and commercial | | 0.9995 |
| Low- to medium-density residential | | 0.9998 |
| Industrial | | 0.9993 |
| Urban green land | | 0.9993 |
| Urban road | | 0.9996 |