API Reference¶

splita exposes 88 classes across 8 modules. Every class returns frozen dataclasses with .to_dict() and a pretty __repr__.

Type legend: Original = algorithm implemented from paper equations. Wrapper = delegates to scipy/sklearn. Hybrid = original algorithm + scipy/sklearn numerical primitives.

Core Analysis (24 classes)¶

Class	Type	Description
`Experiment`	Hybrid	Frequentist A/B test (z-test, t-test, Mann-Whitney, chi-square, delta method, bootstrap)
`BayesianExperiment`	Original	Bayesian A/B test with P(B>A), expected loss, ROPE
`ObjectiveBayesianExperiment`	Original	Empirical Bayes A/B test with prior learned from historical experiments
`QuantileExperiment`	Original	Bootstrap inference at arbitrary quantiles (median, p90, p99)
`SampleSize`	Hybrid	Power analysis for proportions, means, and ratios
`SRMCheck`	Wrapper	Sample Ratio Mismatch detector (chi-square goodness-of-fit)
`MultipleCorrection`	Original	p-value correction (BH, Bonferroni, Holm, BY)
`PermutationTest`	Original	Exact distribution-free hypothesis testing via label permutation
`PowerSimulation`	Wrapper	Monte Carlo power for complex designs
`HTEEstimator`	Wrapper	Heterogeneous treatment effects (T-learner, S-learner)
`CausalForest`	Hybrid	Honest T-learner with jackknife CIs
`TriggeredExperiment`	Wrapper	Intent-to-treat vs per-protocol analysis
`DilutionAnalysis`	Hybrid	Dilute triggered effect back to full ITT population
`InteractionTest`	Hybrid	Segment-level treatment effect heterogeneity (Cochran's Q)
`MultiObjectiveExperiment`	Wrapper	Pareto analysis across multiple metrics
`StratifiedExperiment`	Original	Neyman-style stratified inference
`FunnelExperiment`	Hybrid	Per-step conversion funnel analysis
`InterleavingExperiment`	Hybrid	Team Draft / Balanced interleaving for ranking comparison
`MetricDecomposition`	Wrapper	Decompose metric into additive components and test each
`MixedEffectsExperiment`	Original	Random-intercept model for repeated-measures experiments
`OECBuilder`	Hybrid	Combine multiple metrics into an Overall Evaluation Criterion
`OptimalProxyMetrics`	Original	Learn optimal weighted proxy for a north star metric
`RiskAwareDecision`	Hybrid	Multi-metric constrained decision framework
`SurvivalExperiment`	Hybrid	Kaplan-Meier survival estimation and log-rank test

Variance Reduction (14 classes)¶

Class	Type	Description
`CUPED`	Original	Pre-experiment covariate adjustment
`CUPAC`	Hybrid	ML-predicted covariate adjustment (cross-validated)
`OutlierHandler`	Hybrid	Winsorize, trim, IQR, DBSCAN clustering
`MultivariateCUPED`	Original	Multi-covariate CUPED extension
`RegressionAdjustment`	Original	Lin's OLS with HC2 robust SEs
`AdaptiveWinsorizer`	Original	Grid-search optimal capping thresholds
`DoubleML`	Hybrid	Double/debiased ML for treatment effects
`ClusterBootstrap`	Original	Cluster-level bootstrap for ratio metrics with within-cluster correlation
`InExperimentVR`	Original	Variance reduction using in-experiment control-group covariates
`NonstationaryAdjustment`	Hybrid	Bias-corrected ATE via time-series decomposition
`PostStratification`	Original	Post-experiment stratification with inverse-variance weighting
`PredictionPoweredInference`	Original	Augment small labeled data with ML predictions for valid inference
`RobustMeanEstimator`	Original	Huber M-estimator, Median of Means, Catoni estimator
`TrimmedMeanEstimator`	Original	Robust ATE via symmetric tail trimming

Sequential Testing (7 classes)¶

Class	Type	Description
`mSPRT`	Original	Always-valid p-values via mixture likelihood ratio
`GroupSequential`	Original	Alpha-spending boundaries (OBF, Pocock, Kim-DeMets)
`EValue`	Original	E-value sequential testing
`ConfidenceSequence`	Original	Time-uniform confidence sequences
`EProcess`	Original	Safe testing with e-processes (GRAPA, universal)
`SampleSizeReestimation`	Original	Mid-experiment sample size adjustment via conditional power
`YEASTSequentialTest`	Original	Tuning-free sequential test via Levy's inequality

Bandits (5 classes)¶

Class	Type	Description
`ThompsonSampler`	Original	Multi-armed bandit (Bernoulli, Gaussian, Poisson)
`LinTS`	Original	Linear Thompson Sampling contextual bandit
`LinUCB`	Original	Upper confidence bound contextual bandit
`BayesianStopping`	Original	Stopping rule evaluator for bandits
`OfflineEvaluator`	Hybrid	Offline policy evaluation via IPS and doubly robust estimator

Causal Inference (19 classes)¶

Class	Type	Description
`DifferenceInDifferences`	Hybrid	Classic two-period DiD with parallel trends check
`SyntheticControl`	Hybrid	Weighted donor combination via constrained optimization
`ClusterExperiment`	Hybrid	Cluster-robust inference with ICC
`SwitchbackExperiment`	Hybrid	Time-based switchback design analysis
`SurrogateEstimator`	Wrapper	Short-term to long-term effect prediction
`SurrogateIndex`	Hybrid	Multi-surrogate index with cross-fitting
`InterferenceExperiment`	Original	Network interference with Horvitz-Thompson estimator
`BipartiteExperiment`	Original	Cross-side exposure mapping for two-sided experiments
`ContinuousTreatmentEffect`	Original	Dose-response curve via kernel-weighted local linear regression
`DoublyRobustEstimator`	Hybrid	Augmented IPW (AIPW) with cross-fitting
`DynamicCausalEffect`	Hybrid	Time-varying treatment effects with doubly robust estimation
`EffectTransport`	Original	Transport treatment effects across populations via inverse odds weighting
`GeoExperiment`	Hybrid	Bayesian synthetic control for geo-level marketing experiments
`InstrumentalVariables`	Original	Two-Stage Least Squares (2SLS) for LATE estimation
`MarketplaceExperiment`	Hybrid	Buyer/seller randomization bias analysis for two-sided marketplaces
`MediationAnalysis`	Original	Baron-Kenny mediation with Sobel test for indirect effects (ACME)
`PropensityScoreMatching`	Original	Logistic propensity scores with nearest-neighbor matching
`RegressionDiscontinuity`	Original	Sharp RDD with local linear regression and IK bandwidth
`TMLE`	Hybrid	Targeted Maximum Likelihood Estimation for ATE

Diagnostics (10 classes)¶

Class	Type	Description
`NoveltyCurve`	Original	Rolling-window novelty/primacy effect detection
`AATest`	Wrapper	Validate randomization via simulation
`EffectTimeSeries`	Hybrid	Cumulative treatment effect over time
`MetricSensitivity`	Hybrid	Pre-experiment power estimation from historical data
`VarianceEstimator`	Original	Distributional analysis with A/B-specific recommendations
`NonStationaryDetector`	Original	CUSUM change-point detection on effect series
`CarryoverDetector`	Wrapper	Detect treatment effect leakage into control via pre/post comparison
`FlickerDetector`	Original	Detect users who switched variants mid-experiment
`PHackingDetector`	Hybrid	p-curve analysis for selective reporting detection
`RandomizationValidator`	Hybrid	Covariate balance check via SMD and chi-squared omnibus test

Design (6 classes)¶

Class	Type	Description
`PairwiseDesign`	Original	Mahalanobis distance matched-pair assignment
`AdaptiveEnrichment`	Original	Mid-experiment subgroup selection based on treatment response
`BayesianExperimentOptimizer`	Hybrid	Surrogate-model-based experiment parameter optimization
`BudgetSplitDesign`	Hybrid	Budget-split design to eliminate cannibalization bias
`FractionalFactorialDesign`	Hybrid	Resolution III+ fractional factorial designs for factor screening
`ResponseAdaptiveRandomization`	Original	Dynamic allocation probabilities favouring better-performing arms

Governance (3 classes)¶

Class	Type	Description
`ExperimentRegistry`	Original	In-memory experiment tracking with date filtering
`ConflictDetector`	Original	Overlapping experiment detection (traffic, metric, segment)
`GuardrailMonitor`	Wrapper	Safety metric monitoring with Bonferroni-corrected stopping rules

Top-level functions¶

Function	Description
`auto()`	Auto-select and apply the best variance reduction + analysis
`check()`	Run a suite of data quality checks
`compare()`	Compare multiple experiment results
`diagnose()`	Run diagnostic checks on an experiment
`explain()`	Generate a plain-English explanation of any result
`log()`	Log an experiment result
`load_log()`	Load previously logged results
`meta_analysis()`	Combine results from multiple experiments
`monitor()`	Real-time monitoring dashboard
`notify()`	Send notifications (Slack, email)
`power_report()`	Generate a power analysis report
`report()`	Generate a full experiment report
`serve()`	Serve splita as a REST API
`simulate()`	Monte Carlo power simulation
`to_latex_table()`	Export results to LaTeX
`what_if()`	What-if scenario analysis
`audit_trail()`	Generate an audit trail for an experiment

Plugin system¶

Function	Description
`register_method()`	Register a custom statistical test method
`unregister_method()`	Remove a registered method
`get_method()`	Retrieve a registered method
`list_methods()`	List all registered methods
`clear_methods()`	Clear all registered methods

Datasets¶

Function	Description
`load_ecommerce()`	E-commerce checkout flow experiment (5,000 users, revenue, segments)
`load_marketplace()`	Two-sided marketplace experiment (3,000 buyers, 800 sellers)
`load_subscription()`	SaaS onboarding experiment (2,000 users, survival data)
`load_mobile_app()`	Mobile recommendation engine experiment (4,000 users, sessions, purchases)

Error classes¶

Class	Description
`SplitaError`	Base exception for all splita errors
`ValidationError`	Input validation failure
`InsufficientDataError`	Not enough data for the requested analysis
`NotFittedError`	Called a method that requires fitting first