Assessing the Applicability of Random Forest, Stochastic Gradient Boosted Model, and Extreme Learning Machine Methods to the Quantitative Precipitation Estimation of the Radar Data: A Case Study to Gwangdeoksan Radar, South Korea, in 2018

Figure 5

Plots of rainfall rate estimations versus observations of tested models for precipitation event #1. (a) ZR1-L1 (RMSE: 11.97, R: 0.75). (b) ZR2-L1 (RMSE: 11.91, R: 0.75). (c) ZR3-L1 (RMSE: 11.98, R: 0.75). (d) ZR5-L1 (RMSE: 11.92, R: 0.75). (e) ZR1-L0 (RMSE: 14.59, R: 0.6). (f) ZR2-L0 (RMSE: 14.62, R: 0.6). (g) ZR3-L0 (RMSE: 14.62, R: 0.6). (h) ZR5-L0 (RMSE: 14.66, R: 0.6). (i) RF1 (RMSE: 11.29, R: 0.79). (j) RF2 (RMSE: 11.3, R: 0.79). (k) RF3 (RMSE: 11.01, R: 0.8). (l) RF5 (RMSE: 11.2, R: 0.8). (m) GBM1 (RMSE: 11.05, R: 0.8). (n) GBM2 (RMSE: 11.05, R: 0.8). (o) GBM3 (RMSE: 11.03, R: 0.42). (p) GBM5 (RMSE: 11.06, R: 0.8). (q) ELM1 (RMSE: 10.3, R: 0.83). (r) ELM2 (RMSE: 10.11, R: 0.84). (s) ELM3 (RMSE: 10.21, R: 0.83). (t) ELM5 (RMSE: 10.06, R: 0.84).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)

(m)

(n)

(o)

(p)

(q)

(r)

(s)

(t)