Machine Learning & Choice Models
Statios provides machine learning and choice modeling tools: supervised learning with explainability (Random Forest, Gradient Boosting, SVM, Decision Trees, neural networks); model evaluation (cross-validation, ROC/AUC, confusion matrix, feature importance, SHAP); and conjoint analysis (traditional, choice-based, hierarchical Bayes) for preference modeling.
31 procedures · 8 with documented limitations. Every result is computed by the open Python engine and is exportable to APA tables and reports.
| Procedure | What it does | Status | Key reference |
|---|---|---|---|
| Random Forest | Fit random forest for classification or regression with feature importances and OOB score. | ✅ Full | Breiman (2001) |
| Gradient Boosting | Fit gradient boosted model (XGBoost if available) with learning curves and feature importances. | ✅ Full | Pedregosa et al. (2011) |
| Support Vector Machine | Fit SVM for classification or regression with standardized features and support vector counts. | ✅ Full | Pedregosa et al. (2011) |
| Decision Tree | Fit decision tree with tree rules, feature importances, and depth/leaf metrics. | ✅ Full | Pedregosa et al. (2011) |
| Cross-Validation | K-fold cross-validation with train/test scores and timing metrics across folds. | ✅ Full | Pedregosa et al. (2011) |
| ROC/AUC Analysis | Binary classification ROC curve with AUC and Youden's J optimal threshold. | ✅ Full | Pedregosa et al. (2011) |
| Confusion Matrix | Multi-class confusion matrix with per-class precision/recall/F1, normalizations, and macro/weighted averages. | ✅ Full | Pedregosa et al. (2011) |
| Classification Metrics | Accuracy, precision, recall, F1, Matthews correlation, Cohen's kappa, and log-loss. | ✅ Full | Pedregosa et al. (2011) |
| Feature Importance | Built-in and permutation-based feature importances with standard errors. | ✅ Full | Pedregosa et al. (2011) |
| SHAP Values | SHAP values via TreeExplainer (exact TreeSHAP) or KernelExplainer (model-agnostic) with expected values. | ⚠️ Limited | Pedregosa et al. (2011) |
| SHAP Summary Plot | SHAP summary (beeswarm) plot data with feature importance ranking and per-sample contributions. | ✅ Full | Pedregosa et al. (2011) |
| Train/Test Predict | Fit model on train set and return test predictions, actuals, residuals/probabilities. | ✅ Full | Pedregosa et al. (2011) |
| MLP Classifier | Scikit-learn multilayer perceptron for classification with configurable hidden layers and activation. | ✅ Full | Pedregosa et al. (2011) |
| MLP Regressor | Scikit-learn multilayer perceptron for regression with configurable hidden layers and activation. | ✅ Full | Pedregosa et al. (2011) |
| Deep Neural Network | Keras/TensorFlow deep network for classification or regression with custom architecture, epochs, batch size. | ✅ Full | Pedregosa et al. (2011) |
| Autoencoder | Unsupervised autoencoder for dimensionality reduction and feature learning. | ✅ Full | Pedregosa et al. (2011) |
| CNN for Tabular Data | Convolutional neural network adapted for tabular classification or regression. | ✅ Full | Pedregosa et al. (2011) |
| Neural Network Hyperparameter Tuning | Grid search over neural network hyperparameters with cross-validation. | ✅ Full | Pedregosa et al. (2011) |
| CHAID | Chi-squared automatic interaction detection for tree-based classification. | ⚠️ Limited | Pedregosa et al. (2011) |
| QUEST | Quick, unbiased, efficient statistical tree method for classification. | ⚠️ Limited | Pedregosa et al. (2011) |
| C5.0 | C5.0 decision tree and rule learner for classification. | ⚠️ Limited | Pedregosa et al. (2011) |
| Ensemble Trees Comparison | Side-by-side comparison of multiple tree-based ensemble methods. | ⚠️ Limited | Pedregosa et al. (2011) |
| Tree Export | Export trained tree model in formats suitable for deployment. | ⚠️ Limited | Pedregosa et al. (2011) |
| Compare Trees | Comparative metrics across tree models (accuracy, depth, interpretability). | ⚠️ Limited | Pedregosa et al. (2011) |
| Traditional Conjoint Analysis | Rating-based conjoint via OLS with dummy-coded attributes and part-worths. | ✅ Full | Seabold & Perktold (2010) |
| Choice-Based Conjoint | Aggregate multinomial logit conjoint from choice data with part-worths and hit rate. | ✅ Full | Seabold & Perktold (2010) |
| Hierarchical Bayes CBC | Empirical-Bayes iterated-MAP CBC with individual part-worths (not true HB-MCMC). | ⚠️ Limited | Seabold & Perktold (2010) |
| Market Simulator | Market share simulation given part-worths and scenarios (logit, first-choice, randomized). | ✅ Full | Seabold & Perktold (2010) |
| Attribute Importance | Relative attribute importance from part-worth ranges (% allocation). | ✅ Full | Seabold & Perktold (2010) |
| Conjoint Design | Fractional experimental design generation (full, orthogonal, D-optimal). | ✅ Full | Seabold & Perktold (2010) |
| Conjoint Segmentation | Segment respondents using K-means on individual part-worths. | ✅ Full | Seabold & Perktold (2010) |
Notes & limitations
- SHAP Values — TreeExplainer unavailable in browser (WASM); fallback to KernelExplainer (sampled Shapley values, not exact TreeSHAP).
- CHAID — Stub implementation; returns placeholder.
- QUEST — Stub implementation; returns placeholder.
- C5.0 — Stub implementation; returns placeholder.
- Ensemble Trees Comparison — Stub implementation; returns placeholder.
- Tree Export — Stub implementation; returns placeholder.
- Compare Trees — Stub implementation; returns placeholder.
- Hierarchical Bayes CBC — Empirical-Bayes / EM-style scheme, NOT Gibbs/Metropolis; samples are posterior modes, not draws; heterogeneity SD understates posterior uncertainty.