Variable: fNRec Ranger result Call: ranger(formulaString.lst[[j]], data = dfs, importance = "impurity", write.forest = TRUE, mtry = t.mrfX$bestTune$mtry, num.trees = 500) Type: Regression Number of trees: 500 Sample size: 5 Number of independent variables: 377 Mtry: 12 Target node size: 5 Variable importance mode: impurity OOB prediction error: 845.2204 R squared: -0.2579609 OOB RMSE: 29.073 Variable importance: [,1] AAIavg_GYGA 0 af_agg_30cm_AWCpF23__M_1km 0 af_agg_30cm_PWP__M_1km 0 af_agg_30cm_TAWCpF23__M_1km 0 af_agg_30cm_TAWCpF23mm__M_1km 0 af_agg_30cm_TETAs__M_1km 0 af_agg_ERZD_TAWCpF23mm__M_1km 0 af_BDRICM_T__M_1km 0 af_ERZD__M_1km 0 Al_M_agg30cm_AF_1km 0 ASSDAC3 0 B02CHE3 0 B04CHE3 0 B07CHE3 0 B13CHE3 0 eXtreme Gradient Boosting 5 samples 377 predictors No pre-processing Resampling: Cross-Validated (3 fold, repeated 1 times) Summary of sample sizes: 3, 3, 4 Resampling results across tuning parameters: eta max_depth nrounds RMSE Rsquared 0.3 2 50 19.49471 1 0.3 2 100 19.50235 1 0.3 2 150 19.50235 1 0.3 3 50 19.49471 1 0.3 3 100 19.50235 1 0.3 3 150 19.50235 1 0.3 4 50 19.49471 1 0.3 4 100 19.50235 1 0.3 4 150 19.50235 1 0.3 5 50 19.49471 1 0.3 5 100 19.50235 1 0.3 5 150 19.50235 1 0.3 6 50 19.49471 1 0.3 6 100 19.50235 1 0.3 6 150 19.50235 1 0.3 7 50 19.49471 1 0.3 7 100 19.50235 1 0.3 7 150 19.50235 1 0.3 8 50 19.49471 1 0.3 8 100 19.50235 1 0.3 8 150 19.50235 1 0.4 2 50 19.50223 1 0.4 2 100 19.50235 1 0.4 2 150 19.50235 1 0.4 3 50 19.50223 1 0.4 3 100 19.50235 1 0.4 3 150 19.50235 1 0.4 4 50 19.50223 1 0.4 4 100 19.50235 1 0.4 4 150 19.50235 1 0.4 5 50 19.50223 1 0.4 5 100 19.50235 1 0.4 5 150 19.50235 1 0.4 6 50 19.50223 1 0.4 6 100 19.50235 1 0.4 6 150 19.50235 1 0.4 7 50 19.50223 1 0.4 7 100 19.50235 1 0.4 7 150 19.50235 1 0.4 8 50 19.50223 1 0.4 8 100 19.50235 1 0.4 8 150 19.50235 1 0.5 2 50 19.50230 1 0.5 2 100 19.50230 1 0.5 2 150 19.50230 1 0.5 3 50 19.50230 1 0.5 3 100 19.50230 1 0.5 3 150 19.50230 1 0.5 4 50 19.50230 1 0.5 4 100 19.50230 1 0.5 4 150 19.50230 1 0.5 5 50 19.50230 1 0.5 5 100 19.50230 1 0.5 5 150 19.50230 1 0.5 6 50 19.50230 1 0.5 6 100 19.50230 1 0.5 6 150 19.50230 1 0.5 7 50 19.50230 1 0.5 7 100 19.50230 1 0.5 7 150 19.50230 1 0.5 8 50 19.50230 1 0.5 8 100 19.50230 1 0.5 8 150 19.50230 1 Tuning parameter 'gamma' was held constant at a value of 0 Tuning parameter 'colsample_bytree' was held constant at a value of 0.8 Tuning parameter 'min_child_weight' was held constant at a value of 1 RMSE was used to select the optimal model using the smallest value. The final values used for the model were nrounds = 50, max_depth = 2, eta = 0.3, gamma = 0, colsample_bytree = 0.8 and min_child_weight = 1. RMSE: 19.495 R2: 1 XGBoost variable importance: Feature Gain Cover Frequency 1: AAIavg_GYGA 0.9813152598 0.76923077 0.76923077 2: af_agg_30cm_PWP__M_1km 0.0183546487 0.17948718 0.17948718 3: af_agg_30cm_TAWCpF23mm__M_1km 0.0003300915 0.05128205 0.05128205 4: NA NA NA NA 5: NA NA NA NA 6: NA NA NA NA 7: NA NA NA NA 8: NA NA NA NA 9: NA NA NA NA 10: NA NA NA NA 11: NA NA NA NA 12: NA NA NA NA 13: NA NA NA NA 14: NA NA NA NA 15: NA NA NA NA Ensemble validation RMSE: 19.156 R2: 0.326 -------------------------------------- Variable: fPRec Ranger result Call: ranger(formulaString.lst[[j]], data = dfs, importance = "impurity", write.forest = TRUE, mtry = t.mrfX$bestTune$mtry, num.trees = 500) Type: Regression Number of trees: 500 Sample size: 7 Number of independent variables: 377 Mtry: 14 Target node size: 5 Variable importance mode: impurity OOB prediction error: 39.41709 R squared: 0.4427708 OOB RMSE: 6.278 Variable importance: [,1] Cu_M_agg30cm_AF_1km 7.596656 Rice_actual_baseline 7.115657 NCluster_19_AF_1km 6.973525 GIEMSD3 6.717234 L15USG5 6.056279 PHIHOXagg0_30 5.670028 C01GLC5 5.636855 NCluster_18_AF_1km 5.464765 M13NDVIALT 5.412783 Temperature 5.397485 NCluster_11_AF_1km 5.228973 VBFMRG5 5.216193 M13RB3A08 5.141460 M43BSALT 5.008347 M13RB1A08 4.903323 eXtreme Gradient Boosting 7 samples 377 predictors No pre-processing Resampling: Cross-Validated (3 fold, repeated 1 times) Summary of sample sizes: 4, 5, 5 Resampling results across tuning parameters: eta max_depth nrounds RMSE Rsquared 0.3 2 50 6.649553 0.5060850 0.3 2 100 6.651727 0.5060849 0.3 2 150 6.651727 0.5060849 0.3 3 50 6.094152 0.5363004 0.3 3 100 6.096352 0.5362997 0.3 3 150 6.096352 0.5362997 0.3 4 50 6.380274 0.5160044 0.3 4 100 6.382477 0.5160042 0.3 4 150 6.382477 0.5160042 0.3 5 50 6.565788 0.5102586 0.3 5 100 6.567987 0.5102585 0.3 5 150 6.567987 0.5102585 0.3 6 50 6.534999 0.5079430 0.3 6 100 6.537204 0.5079429 0.3 6 150 6.537204 0.5079429 0.3 7 50 6.850850 0.5032205 0.3 7 100 6.853017 0.5032215 0.3 7 150 6.853017 0.5032215 0.3 8 50 6.860977 0.5101682 0.3 8 100 6.863185 0.5101680 0.3 8 150 6.863185 0.5101680 0.4 2 50 6.786967 0.5038868 0.4 2 100 6.786977 0.5038868 0.4 2 150 6.786977 0.5038868 0.4 3 50 6.484035 0.5102966 0.4 3 100 6.484045 0.5102966 0.4 3 150 6.484045 0.5102966 0.4 4 50 6.287042 0.5333912 0.4 4 100 6.287052 0.5333912 0.4 4 150 6.287052 0.5333912 0.4 5 50 6.260953 0.5225646 0.4 5 100 6.260964 0.5225646 0.4 5 150 6.260964 0.5225646 0.4 6 50 6.397649 0.5142234 0.4 6 100 6.397659 0.5142234 0.4 6 150 6.397659 0.5142234 0.4 7 50 6.467638 0.5110883 0.4 7 100 6.467649 0.5110883 0.4 7 150 6.467649 0.5110883 0.4 8 50 6.680104 0.5046395 0.4 8 100 6.680115 0.5046395 0.4 8 150 6.680115 0.5046395 0.5 2 50 6.088551 0.5366694 0.5 2 100 6.088551 0.5366694 0.5 2 150 6.088551 0.5366694 0.5 3 50 6.530669 0.5081523 0.5 3 100 6.530669 0.5081523 0.5 3 150 6.530669 0.5081523 0.5 4 50 6.833020 0.5026969 0.5 4 100 6.833020 0.5026969 0.5 4 150 6.833020 0.5026969 0.5 5 50 5.803973 0.5690488 0.5 5 100 5.803973 0.5690488 0.5 5 150 5.803973 0.5690488 0.5 6 50 6.109341 0.5345860 0.5 6 100 6.109341 0.5345860 0.5 6 150 6.109341 0.5345860 0.5 7 50 7.004320 0.5026834 0.5 7 100 7.004320 0.5026834 0.5 7 150 7.004320 0.5026834 0.5 8 50 6.710181 0.5028840 0.5 8 100 6.710181 0.5028840 0.5 8 150 6.710181 0.5028840 Tuning parameter 'gamma' was held constant at a value of 0 Tuning parameter 'colsample_bytree' was held constant at a value of 0.8 Tuning parameter 'min_child_weight' was held constant at a value of 1 RMSE was used to select the optimal model using the smallest value. The final values used for the model were nrounds = 50, max_depth = 5, eta = 0.5, gamma = 0, colsample_bytree = 0.8 and min_child_weight = 1. RMSE: 5.804 R2: 0.569 XGBoost variable importance: Feature Gain Cover Frequency 1: ASSDAC3 0.4958845681 0.03398058 0.02857143 2: af_agg_30cm_PWP__M_1km 0.3507341521 0.51941748 0.45714286 3: B13CHE3 0.1141189251 0.10194175 0.08571429 4: af_agg_ERZD_TAWCpF23mm__M_1km 0.0213231635 0.06796117 0.05714286 5: AAIavg_GYGA 0.0139069207 0.18932039 0.28571429 6: af_agg_30cm_TAWCpF23mm__M_1km 0.0039198099 0.01941748 0.02857143 7: af_BDRICM_T__M_1km 0.0001124606 0.06796117 0.05714286 8: NA NA NA NA 9: NA NA NA NA 10: NA NA NA NA 11: NA NA NA NA 12: NA NA NA NA 13: NA NA NA NA 14: NA NA NA NA 15: NA NA NA NA Ensemble validation RMSE: 6.408 R2: 0.396 -------------------------------------- Variable: fKRec Ranger result Call: ranger(formulaString.lst[[j]], data = dfs, importance = "impurity", write.forest = TRUE, mtry = t.mrfX$bestTune$mtry, num.trees = 500) Type: Regression Number of trees: 500 Sample size: 6 Number of independent variables: 377 Mtry: 18 Target node size: 5 Variable importance mode: impurity OOB prediction error: 29.05159 R squared: 0.581282 OOB RMSE: 5.39 Variable importance: [,1] M13RB1A08 8.591536 REDL14 5.946238 M43BVALT 5.613886 M13RB3A01 4.640969 NCluster_20_AF_1km 4.628003 NCluster_11_AF_1km 4.592571 Slopeclassc1 4.491019 EXMOD5avg 4.447985 NIRL14 4.431493 M13NDVIA08 4.334838 Cu_M_agg30cm_AF_1km 4.296405 L10USG5 4.123293 Slopeclassc2 4.110323 M13RB1ALT 4.040611 yFertilised_RiceTrials 4.024667 eXtreme Gradient Boosting 6 samples 377 predictors No pre-processing Resampling: Cross-Validated (3 fold, repeated 1 times) Summary of sample sizes: 4, 4, 4 Resampling results across tuning parameters: eta max_depth nrounds RMSE Rsquared 0.3 2 50 4.471252 1 0.3 2 100 4.467755 1 0.3 2 150 4.467755 1 0.3 3 50 4.471252 1 0.3 3 100 4.467755 1 0.3 3 150 4.467755 1 0.3 4 50 4.471252 1 0.3 4 100 4.467755 1 0.3 4 150 4.467755 1 0.3 5 50 4.471252 1 0.3 5 100 4.467755 1 0.3 5 150 4.467755 1 0.3 6 50 4.471252 1 0.3 6 100 4.467755 1 0.3 6 150 4.467755 1 0.3 7 50 4.471252 1 0.3 7 100 4.467755 1 0.3 7 150 4.467755 1 0.3 8 50 4.471252 1 0.3 8 100 4.467755 1 0.3 8 150 4.467755 1 0.4 2 50 4.468001 1 0.4 2 100 4.467892 1 0.4 2 150 4.467892 1 0.4 3 50 4.468001 1 0.4 3 100 4.467892 1 0.4 3 150 4.467892 1 0.4 4 50 4.468001 1 0.4 4 100 4.467892 1 0.4 4 150 4.467892 1 0.4 5 50 4.468001 1 0.4 5 100 4.467892 1 0.4 5 150 4.467892 1 0.4 6 50 4.468001 1 0.4 6 100 4.467892 1 0.4 6 150 4.467892 1 0.4 7 50 4.468001 1 0.4 7 100 4.467892 1 0.4 7 150 4.467892 1 0.4 8 50 4.468001 1 0.4 8 100 4.467892 1 0.4 8 150 4.467892 1 0.5 2 50 4.467826 1 0.5 2 100 4.467826 1 0.5 2 150 4.467826 1 0.5 3 50 4.467826 1 0.5 3 100 4.467826 1 0.5 3 150 4.467826 1 0.5 4 50 4.467826 1 0.5 4 100 4.467826 1 0.5 4 150 4.467826 1 0.5 5 50 4.467826 1 0.5 5 100 4.467826 1 0.5 5 150 4.467826 1 0.5 6 50 4.467826 1 0.5 6 100 4.467826 1 0.5 6 150 4.467826 1 0.5 7 50 4.467826 1 0.5 7 100 4.467826 1 0.5 7 150 4.467826 1 0.5 8 50 4.467826 1 0.5 8 100 4.467826 1 0.5 8 150 4.467826 1 Tuning parameter 'gamma' was held constant at a value of 0 Tuning parameter 'colsample_bytree' was held constant at a value of 0.8 Tuning parameter 'min_child_weight' was held constant at a value of 1 RMSE was used to select the optimal model using the smallest value. The final values used for the model were nrounds = 100, max_depth = 2, eta = 0.3, gamma = 0, colsample_bytree = 0.8 and min_child_weight = 1. RMSE: 4.468 R2: 1 XGBoost variable importance: Feature Gain Cover Frequency 1: af_BDRICM_T__M_1km 0.6389745966 0.12820513 0.12820513 2: af_agg_30cm_PWP__M_1km 0.3607005376 0.84615385 0.84615385 3: AAIavg_GYGA 0.0003248657 0.02564103 0.02564103 4: NA NA NA NA 5: NA NA NA NA 6: NA NA NA NA 7: NA NA NA NA 8: NA NA NA NA 9: NA NA NA NA 10: NA NA NA NA 11: NA NA NA NA 12: NA NA NA NA 13: NA NA NA NA 14: NA NA NA NA 15: NA NA NA NA Ensemble validation RMSE: 5.756 R2: 0.544 --------------------------------------