Package index • oda

Datasets

Built-in example datasets used in vignettes and canon fixture tests.

cta_demo: CTA demonstration dataset
myeloma: Myeloma gene-expression dataset (CTA benchmark)

ODA - fitting and prediction

Unified dispatcher, binary-class engine, multiclass engine, rule application, and performance metrics.

oda_fit(): Fit an ODA model
oda_univariate_core(): Fit a univariate binary-class ODA model
oda_multiclass_unioda_core(): Fit a univariate multiclass ODA model
oda_rule_predict(): Apply a binary ODA rule to new data
oda_rule_predict_multiclass(): Apply a multiclass ODA rule to new data
oda_propensity_weights(): ODA rule strata propensity weights
oda_power(): ODA power analysis via simulation
oda_sample_size(): ODA minimum sample size via bisection

ODA - S3 methods and accessors

S3 methods for oda_fit objects; accessor functions for predictions, confusion, metrics, and D statistic.

predict(<oda_fit>): Predict class labels from a fitted ODA model
print(<oda_fit>): Print a fitted ODA model
summary(<oda_fit>): Summarize a fitted ODA model
oda_predictions(): Retrieve predictions from a fitted ODA model
oda_confusion(): Retrieve a confusion matrix from a fitted ODA model
oda_metrics(): Retrieve scalar performance metrics from a fitted ODA model
oda_d_stat(): Compute the D statistic for a fitted ODA model

ODA - confusion and performance helpers

oda_confusion_binary(): Binary confusion table
oda_confusion_multiclass(): Multiclass confusion matrix
oda_mean_pac(): Mean PAC from sensitivity and specificity
oda_ess_from_meanpac(): Effect Strength for Sensitivity from mean PAC
oda_ess_from_mean(): ESS from mean metric for a C-class problem

ODA - data preparation and validation

Preprocessing, attribute-type inference, group and weight validation, and readiness checking.

oda_clean_missing_codes(): Replace missing-code values with NA
oda_infer_attr_types(): Infer attribute types from a predictor data frame
oda_validate_group(): Validate a class / group variable
oda_validate_weights(): Validate a case weight vector
oda_readiness_check(): Preflight readiness check for ODA / CTA analysis

CTA - fitting and prediction

Classification Tree Analysis: public entry point, internal engine, prediction, and MDSA family fitting.

cta_fit(): Fit a Classification Tree Analysis (CTA) model (public wrapper)
oda_cta_fit(): Fit a Classification Tree Analysis (CTA) model (internal engine)
predict(<cta_tree>): Classify new observations using a CTA tree
print(<cta_tree>): Print a CTA tree in MegaODA node table format
summary(<cta_tree>): Summarize a fitted CTA tree
cta_descendant_family(): MDSA descendant family for CTA

CTA - tree structure

cta_node_table(): Canonical CTA node report table
cta_strata(): Number of terminal leaf endpoints in a CTA tree
cta_endpoint_denominators(): Terminal endpoint denominators of a CTA tree
cta_min_terminal_denom(): Minimum terminal endpoint denominator of a CTA tree
cta_d_stat(): D statistic for a fitted CTA tree

CTA - reporting and translation

On-demand translation functions for endpoints, staging, propensity weights, observation weights, and confusion tables. All computed on explicit call; nothing extra is stored at fit time (lean-fit invariant).

cta_endpoint_summary(): Endpoint reporting summary for a fitted CTA tree
cta_endpoint_counts(): Per-endpoint class count table for a fitted CTA tree
cta_endpoint_table(): Canonical terminal endpoint map for a fitted CTA tree
cta_staging_table(): Staging table for a fitted CTA tree
cta_propensity_weights(): Endpoint-level propensity-score weights for a fitted CTA tree
cta_assign_endpoints(): Assign observations to CTA terminal endpoints
cta_observation_weights(): Assign per-observation CTA propensity weights
cta_confusion_table(): Final selected tree training confusion table
as_confusion_matrix(): Convert a tidy confusion data frame to a 2x2 integer matrix
cta_confusion_matrix(): Extract training confusion matrix from a fitted CTA tree

MDSA - family summary

MINDENOM descendant family: sequence of progressively simpler trees, D-statistic complexity measure, and minimum-D model selection.

cta_family_table(): Tidy table of a CTA descendant family
summary(<cta_family>): Summarise a CTA descendant family
print(<cta_family>): Print a CTA descendant family
print(<cta_family_summary>): Print a CTA family summary

LORT - Locally Optimal Recursive Trees

Greedy recursive CTA: LORT builds a sequence of CTA models, one per stratum of a prior tree. lort_fit() is the convenience entry point; cta_fit(recursive = TRUE) is the equivalent internal path.

lort_fit(): Fit a Locally Optimal Recursive Tree (LORT)
predict(<cta_ort>): Predict method for Locally Optimal Recursive Tree (LORT)
print(<cta_ort>): Print method for Locally Optimal Recursive Tree (LORT)
summary(<cta_ort>): Summary method for Locally Optimal Recursive Tree (LORT)
plot(<cta_ort>): Plot method for Locally Optimal Recursive Tree (LORT)
cta_ort_node_table(): Node-level summary table for a fitted LORT (legacy name: cta_ort)
ort_plot_data(): Renderer-independent layout data for a LORT composite tree
lort_index_path(): LORT path from root to a given node index
lort_local_tree(): Extract the local CTA model embedded at a LORT node
lort_path_table(): Formatted path table for a LORT recursion path
lort_propensity_weights(): LORT terminal strata propensity weights

SDA - Sequential Discriminant Analysis

SDA fits a sequence of ODA models, removing correctly classified observations after each step. auto_sda_plan() produces a staged workflow plan; helpers convert SDA output for downstream CTA use.

sda_fit(): Run a Structural Decomposition Analysis (SDA) procedure
predict(<sda_fit>): Predict from an SDA procedure result
print(<sda_fit>): Print an sda_fit object
summary(<sda_fit>): Summarise an sda_fit object
auto_sda_plan(): Dry-run planning and validation layer for SDA
print(<auto_sda_plan>): Print an auto_sda_plan object
sda_candidate_table(): Return the candidate table from one or all SDA steps
sda_selected_attributes(): Return the selected attribute names from an SDA procedure result
sda_step_table(): Return a summary table of SDA steps
sda_to_cta_data(): Prepare X and y for CTA using SDA-selected attributes
as_cta_candidates(): Subset a data frame to the SDA-selected candidate columns
sda_anchor(): Construct an sda_anchor object
as_sda_anchor(): Convert an object to an sda_anchor
validate_sda_anchor(): Validate an sda_anchor object
print(<sda_anchor>): Print an sda_anchor
summary(<sda_anchor>): Summarise an sda_anchor

Covariate balance diagnostics

Univariate ODA balance table, SMD companion, CTA multivariate balance table, and plot-data transforms. All balance functions treat the outcome variable as out of scope; group (treatment/exposure) is the classifier.

oda_balance_table(): Univariate ODA covariate balance diagnostics
smd_balance_table(): Conventional SMD companion table for covariate balance
oda_balance_plot_data(): Renderer-ready plot data for univariate ODA covariate balance
cta_balance_table(): Multivariate CTA covariate balance diagnostics
cta_balance_plot_data(): Renderer-ready plot data for CTA covariate balance
oda_balance_effect_table(): ODA covariate balance evidence-interval table
cta_balance_effect_summary(): CTA covariate balance evidence-interval summary
propensity_ess_balance(): Propensity-weighted ESS balance diagnostic

Graphics - base-R renderers

cta_plot_data() and ort_plot_data() produce the renderer-independent tree diagram data contract. plot.cta_tree() and plot.cta_ort() are the native base-R renderers.

cta_plot_data(): Extract layout data for plotting a CTA tree
ort_plot_data(): Renderer-independent layout data for a LORT composite tree
plot(<cta_tree>): Plot a fitted CTA tree
plot(<cta_ort>): Plot method for Locally Optimal Recursive Tree (LORT)

Graphics v3 - ggplot2 renderers

ggplot2-based tree diagrams and balance plots. Requires ggplot2 (in Suggests). All six functions are pure renderers: they never call fitting functions. Pass pre-computed plot-data or the fit object directly.

plot_cta_tree(): Plot a CTA tree using ggplot2
plot_lort_tree(): Plot a LORT (Locally Optimal Recursive Tree) using ggplot2
plot_lort_path(): Plot the full local CTA models along a LORT recursion path
plot_cta_family(): Plot a CTA descendant family member using ggplot2
plot_oda_balance(): Plot ODA covariate balance
plot_smd_balance(): Plot SMD covariate balance
plot_balance_love(): Love plot for covariate balance (SMD)
plot_cta_balance(): Plot CTA multivariate covariate balance
plot_oda_balance_effects(): Forest plot of ODA covariate balance evidence intervals
plot_cta_balance_effects(): Evidence card for CTA multivariate covariate balance

NOVOboot

Fixed-confusion NOVOmetric bootstrap confidence intervals.

novo_boot_ci() print(<novo_boot_ci>): Novometric bootstrap CI from a fixed 2x2 confusion matrix

Developer / internal helpers

Low-level engine helpers and secondary print methods. These are not part of the primary user-facing API; documented for developer reference only.

.lort_parent_maps(): Build parent map and endpoint-index map for LORT nodes (internal helper used by lort_index_path and lort_path_table)
oda_best_ordered_multiclass_partition(): Select the best K-segment ordered partition by MegaODA spec: PRIMARY -> SECONDARY -> FIRST IDENTIFIED (enum order via tick()).
oda_loo_multiclass_ordered(): Leave-one-out cross-validation for ordered multiclass ODA.
oda_mc_p_value(): Monte Carlo Fisher-randomization p-value with Clopper-Pearson early stopping.
print(<oda_fit_summary>): Print an ODA fit summary
print(<cta_tree_summary>): Print a CTA tree summary
print(<cta_ort_summary>): Print method for cta_ort_summary
print(<sda_fit_summary>): Print an sda_fit_summary object