rebase with paper/benchmarks and finalizing benchmakrs

Author: vyron-arvanitis

paper/paper.md

@@ -8,8 +8,10 @@ tags:
   - uncertainty quantification
 authors:
   - name: Vyron Arvanitis
+    orcid: 0009-0001-2290-5084
     equal-contrib: true
     affiliation: 1
+    email: vyronas.arvanitis@gmail.com
   - name: Angelos Aslanidis
     orcid: 0009-0009-6699-2691
     equal-contrib: true
@@ -82,11 +84,16 @@ regressor = BdeRegressor(
 regressor.fit(x=X_train, y=y_train)
 
 means, sigmas = regressor.predict(X_test, mean_and_std=True)
-mean, intervals = regressor.predict(X_test, credible_intervals=[0.1, 0.9])
+means, intervals = regressor.predict(X_test, credible_intervals=[0.1, 0.9])
 raw = regressor.predict(X_test, raw=True)
 ```
+
 Classification follows analogously using `BdeClassifier`.
 
+# Benchmark for Regression
+
+<!-- TBD -->
+
 # AI usage
 
 Generative AI was used for smart code autocompletion via GitHub Copilot, using the Claude Sonnet 3.7 and Claude Sonnet 4 models. Its use was limited to local, line- or block-level completion during software development. Further Codex was used to assist in the generation of the first draft of the package documentation and for few refactors ensuring the compatibility with the `scikit-learn` API. No AI tools were used for ideation, architectural or design decisions, code review or testing strategy. All generated suggestions were critically reviewed, modified where necessary, and fully validated by the authors, who retain complete responsibility for the correctness, originality, licensing compliance, and ethical integrity of all materials.

scripts/airfoil/architecture.txt

@@ -28,6 +28,7 @@ xgb.XGBRegressor(
         tree_method="hist",
         n_estimators=2000,
         learning_rate=0.05,
+        subsample=0.8,
         max_depth=6,
         random_state=seed,
     )

scripts/airfoil/results.csv

@@ -4,11 +4,11 @@ airfoil,rf,1,0.2629648316900271,0.08320355764922648,0.9314174088649452
 airfoil,rf,2,0.2626142321588769,0.08186941172420922,0.932800805196166
 airfoil,rf,3,0.2651418412585877,0.09144818704596017,0.9378628369886428
 airfoil,rf,4,0.2706200458877758,0.11189904662337234,0.9315889999270439
-airfoil,xgb,0,0.201022207736969,-0.18540139712257897,4.04983551800251
-airfoil,xgb,1,0.201022207736969,-0.18540139712257897,3.3374153377953917
-airfoil,xgb,2,0.201022207736969,-0.18540139712257897,2.934643105836585
-airfoil,xgb,3,0.201022207736969,-0.18540139712257897,3.2989675868302584
-airfoil,xgb,4,0.201022207736969,-0.18540139712257897,3.430911511881277
+airfoil,xgb,0,0.1911470890045166,-0.23577354200568518,1.0079011999769136
+airfoil,xgb,1,0.19747740030288696,-0.20319265487405683,1.0418417999753729
+airfoil,xgb,2,0.19886420667171478,-0.19619454380656484,1.0558768000919372
+airfoil,xgb,3,0.20475415885448456,-0.16700672113076545,1.0535908000310883
+airfoil,xgb,4,0.20543842017650604,-0.1636705030501022,1.0588497000280768
 airfoil,linear,0,0.6867813467979431,1.0431992708559026,0.0020498689264059067
 airfoil,linear,1,0.6867813467979431,1.0431992708559026,0.000771824037656188
 airfoil,linear,2,0.6867813467979431,1.0431992708559026,0.00073277298361063