scikit-learn API (XGBoost >= 2)
import xgboost as xgb
clf = xgb.XGBClassifier(
objective="binary:logistic", eval_metric=["aucpr", "logloss"], tree_method="hist",
learning_rate=0.05, n_estimators=10000, max_depth=6, min_child_weight=5,
subsample=0.8, colsample_bytree=0.8, reg_lambda=1.0, reg_alpha=0.0, gamma=0.0,
early_stopping_rounds=200, random_state=0, n_jobs=-1)
clf.fit(X_tr, y_tr, eval_set=[(X_va, y_va)], verbose=200)
print(clf.best_iteration, clf.best_score)
proba = clf.predict_proba(X_te)[:, 1] # uses best_iteration automatically when early stopping was onIn XGBoost 2+/3, early_stopping_rounds, eval_metric and callbacks belong in the constructor, not in
fit(). The last metric in eval_metric drives early stopping.
Native API
dtrain = xgb.DMatrix(X_tr, label=y_tr, enable_categorical=True)
dvalid = xgb.DMatrix(X_va, label=y_va, enable_categorical=True)
booster = xgb.train(params, dtrain, num_boost_round=10000, evals=[(dtrain, "train"), (dvalid, "valid")],
early_stopping_rounds=200, verbose_eval=200)
cv = xgb.cv(params, dtrain, num_boost_round=10000, nfold=5, stratified=True, early_stopping_rounds=200, seed=0)Key parameters
| Parameter | Effect |
|---|---|
max_depth | depth-wise growth capacity (default 6) |
grow_policy="lossguide" + max_leaves | leaf-wise growth like LightGBM |
min_child_weight | minimum hessian sum per child (≈ rows for squared error; p(1-p)·rows for logloss) |
gamma (min_split_loss) | minimum loss reduction to split |
subsample, colsample_bytree / bylevel / bynode | sampling |
reg_lambda (default 1), reg_alpha | L2 / L1 on leaf weights |
max_delta_step | caps leaf updates; helps extreme imbalance with logloss |
scale_pos_weight | weight for positives (≈ negatives/positives) |
tree_method="hist" | default fast histogram method; device="cuda" for GPU |
max_cat_to_onehot | one-hot style splits for categoricals with few levels |
monotone_constraints, interaction_constraints | domain constraints |
multi_strategy="multi_output_tree" | one tree for all outputs (multi-target) |
Categoricals
enable_categorical=True with pandas category dtype (or DMatrix feature types). Keep identical category
sets between train and inference.
Objectives and metrics
binary:logistic, multi:softprob (num_class), reg:squarederror, reg:absoluteerror,
reg:pseudohubererror, reg:quantileerror (quantile_alpha, multi-quantile), count:poisson, reg:gamma,
reg:tweedie, rank:ndcg / rank:pairwise, survival:cox, survival:aft. Metrics: auc, aucpr,
logloss, error, rmse, mae, ndcg@k, map@k.
Common problems
| Problem | Fix |
|---|---|
TypeError: fit() got an unexpected keyword argument 'early_stopping_rounds' | move it to the constructor (XGBoost 2+) |
| Categorical error with object dtype | convert to category and set enable_categorical=True |
Probabilities poorly calibrated after scale_pos_weight | expected: recalibrate or do not weight |
| Very slow on wide sparse data | tree_method="hist", colsample_bytree < 1, sparse CSR input |
| Different results on GPU vs CPU | small numeric differences are normal; fix seeds and versions |
anomaly-detection-model-builder
Build and evaluate anomaly detection for tabular, sensor and transaction data with few or no labels: define the anomaly type (point, contextual, collective), choose detectors (robust z-score/MAD and IQR rules, Mahalanobis with robust covariance, k-NN distance, LOF, Isolation Forest, one-class SVM, PCA reconstruction / autoencoders, forecasting residuals for time series), combine them by rank, explain each alert by its deviating features, set thresholds from an alert budget or the score distribution, and evaluate with precision@k, PR-AUC and recall on known incidents, or detector agreement when no labels exist. Includes pure-Python detectors (robust z, Mahalanobis, kNN, Isolation Forest, ensemble) and an evaluation script.
catboost-model-builder
Build CatBoost models that use its strengths correctly: raw categorical columns as cat_features (ordered target statistics, no manual encoding), text_features, Pool construction with NaN-safe categories, ordered boosting and has_time for temporal data, symmetric (oblivious) trees, overfitting detector with eval_set and use_best_model, losses for classification, regression (Quantile, Poisson, Tweedie) and ranking (YetiRank), class weights, monotone constraints, feature importance (PredictionValuesChange, LossFunctionChange) and SHAP values, GPU training and model export. Plans cat/text/numeric columns from a CSV and generates a runnable training script, and checks parameter sets for the conflicts CatBoost rejects (iterations vs n_estimators, Bayesian bootstrap with subsample, multiple class-weight options, depth > 16, use_best_model without eval_set).
categorical-encoding-selector
Choose and implement the right encoding for every categorical column given the model family and the data: one-hot (with infrequent-level grouping and handle_unknown), ordinal with an explicit order, cross-fitted target encoding (scikit-learn TargetEncoder, CatBoost ordered statistics, category_encoders inside CV), frequency/count encoding, feature hashing, native categorical support in LightGBM, HistGradientBoosting, XGBoost and CatBoost, prefix levels for codes (zip3, NAICS-2), and embeddings. Measures cardinality, rare-level share, unseen-level rate on a time-ordered holdout and target signal per column, emits the ColumnTransformer, and quantifies target-encoding leakage (in-sample vs out-of-fold vs holdout AUC).
classification-model-selector
Choose the right classification algorithm for tabular, text or mixed data and prove the choice with a fair bake-off: logistic regression (L2, L1, elastic net), linear SVM, naive Bayes, k-NN, RBF SVM, decision trees, random forest, extra trees, HistGradientBoosting, LightGBM, XGBoost, CatBoost, explainable boosting machines (EBM/GAM), points scorecards and MLPs, matched to data size, width, categorical and high-cardinality share, missingness, class imbalance, text, and constraints such as interpretability, monotonic constraints, calibrated probabilities, latency budgets and tiny deployable models. Ranks candidates with starting hyperparameters and cautions, then generates a bake-off script running every candidate on identical CV folds with per-fold metrics, fit time and prediction latency.
clustering-algorithm-selector
Choose the right clustering algorithm and check that clustering makes sense at all: cluster tendency (Hopkins statistic with a correlated-features caveat, nearest-neighbour isolation), then k-means / MiniBatchKMeans, Gaussian mixtures with BIC, agglomerative (Ward, average, complete, single) with dendrograms, DBSCAN, HDBSCAN, OPTICS, spectral clustering, BIRCH, mean shift, k-prototypes and k-modes for mixed or categorical data, and Gower distance, matched to data size, expected shapes, noise, varying density, known k, soft memberships, hierarchy and streaming. Covers preprocessing (scaling, encoding, dimensionality, distance metrics), validation (internal indices, bootstrap stability, external labels, business profiling) and assigning new points.
collaborative-filtering-implementer
Implement collaborative filtering recommenders that beat popularity: item-based kNN with cosine, Jaccard or asymmetric similarity, shrinkage, neighbourhood size, positive-only or mean-centred rating signals and 'because you liked X' explanations; implicit ALS with confidence weighting, BPR pairwise ranking and the closed-form EASE model in plain NumPy, tuned on a temporal validation split by recall@k and NDCG@k and refit on all data; plus user-kNN, explicit matrix factorisation, the implicit/LightFM libraries, hyperparameter ranges, cold-start fallbacks, scaling with sparse matrices and approximate nearest neighbours. Produces top-k lists in the format of the recommendation evaluator.