124
K. Shrestha et al.
Fig. 7.10 Distribution of
landslide with drainage
density
The whole model test shows that the-loglikelihood has dropped to 30.59 (from
68.62) and the difference in model by including the variable is significant with
Prob > ChiSq (<0.0001) and chi-square statistics (76.05009). This means models
are compared and hypothesis is tested which showed model improvement has been
achieved.
The result of landslide susceptibility index classification showed the maximum
value of 0.7309 and the minimum value of 0.5. The result obtained showed that about
25% of the landslide has the LSI less than 0.62, about 50% of landslide is in the range
of 0.62–0.64, about 75% of landslide had at the range of 0.64–0.645 and 100% of
landslide had the range of 0.645–0.7309 (Figs. 7.18 and 7.19, Table 7.5).
7.13 Comparing LSI from Both: Receiver Operating
Characteristic (ROC)
Landslide Susceptibility Models after the preparation should be validated before
using to ascertain their reliability. According to Vakhshoori and Zare (2018), the
validity of LSMs had simply been presented by simple statistics such as landslide
K. Shrestha et al.
Fig. 7.10 Distribution of
landslide with drainage
density
The whole model test shows that the-loglikelihood has dropped to 30.59 (from
68.62) and the difference in model by including the variable is significant with
Prob > ChiSq (<0.0001) and chi-square statistics (76.05009). This means models
are compared and hypothesis is tested which showed model improvement has been
achieved.
The result of landslide susceptibility index classification showed the maximum
value of 0.7309 and the minimum value of 0.5. The result obtained showed that about
25% of the landslide has the LSI less than 0.62, about 50% of landslide is in the range
of 0.62–0.64, about 75% of landslide had at the range of 0.64–0.645 and 100% of
landslide had the range of 0.645–0.7309 (Figs. 7.18 and 7.19, Table 7.5).
7.13 Comparing LSI from Both: Receiver Operating
Characteristic (ROC)
Landslide Susceptibility Models after the preparation should be validated before
using to ascertain their reliability. According to Vakhshoori and Zare (2018), the
validity of LSMs had simply been presented by simple statistics such as landslide