Grid Prediction - Prediction Engine

<div class="section-label">02 Grid Prediction</div> <div class="section-title">The primary form of RBP analysis.</div> Evaluates combinations of predictive attributes across relevance and similarity thresholds, then forms a fit-weighted composite prediction for circumstances theta. Also returns variable-level **Impact on Fit** and **Impact on Prediction**. Recommended default path for analysis. Same input shapes as [[Predict]]. ## Signature ```text predict_grid( y, # outcomes, length N X, # attributes, N × K theta, # circumstances, length K options? # GridOptions; omit → defaults ) → results ``` **Call:** `predict_grid` **Options:** `GridOptions` (PredictOptions base + Grid fields) **Returns:** [[Results/Overview|Results]] (composite; typically T = 1) ## Arguments | Name | Shape | Description | | --- | --- | --- | | y | N | Observed outcomes | | X | N × K | Training attributes aligned with y | | theta | K | Circumstances for the prediction task | | options | (object) | GridOptions; optional, omit for defaults | N = number of observations (rows), K = number of independent variables (columns). ## Options Omit options to use Grid defaults. Full field list: [[Config/Options#GridOptions|GridOptions]]. Allowed value names: [[Config/Allowed Values|Allowed values]]. ### Grid fields | Name | Type / values | Default | Description | | --- | --- | --- | --- | | max_iter | integer > 0 | 1000 | Cap on combination search iterations | | k | integer > 0 | 1 | Combination size / sampling parameter | | seed | unsigned integer | 42 | RNG seed for combination sampling | | attribute_combi | matrix Q × K | generated | Fixed combination matrix; omit to sample | | retain_grid_objects | omit \| comma-separated keys | omit (lean) | Which per-cell objects to keep | | inner_parallel | auto \| off | auto | Parallel over combinations | ### Retain keys yhat_cells, adjusted_fit_cells, n_cells, weights_cells, k_cells, combi_cells, ysolo_distribution Name only the keys you need. Descriptions and shapes: [[Config/Options#Retain grid objects|Retain grid objects]] · [[Results/Grid Cells|Grid Cells]]. ### Shared base (inherited) Includes threshold, censor_type, censor_unit, censor_operator, prediction_scale, adj_fit_multiplier, inv_method, and flags. See [[Config/Options#PredictOptions|PredictOptions]] and [[Config/Options#GridOptions|GridOptions]]. | Notable Grid defaults | Value | | --- | --- | | threshold | [0, 0.2, 0.5, 0.8] | | censor_type | both | ## Results Composite handle. Typically T = 1. Full field map: [[Results/Overview|Results]]. | Present | Fields | Detail | | --- | --- | --- | | Always (success) | Root core: yhat, fit, adjusted_fit, … | [[Results/Core\|Core]]. Scalar forecast: float(result.yhat[0]) (Python) | | Usual (retain off) | grid_insights: **Impact on Fit**, **Impact on Prediction**, … | [[Results/Grid Insights\|Grid Insights]] | | If retain set | grid_cells (plus solo histogram if ysolo_distribution retained) | [[Results/Grid Cells\|Grid Cells]] | prediction_weights, auxiliary, and solo_distribution are often empty on the Grid composite compared with [[Predict]]. <div class="btn-row"> <a class="btn-primary" href="/Results/Grid%20Insights">Grid Insights</a> <a class="btn-ghost" href="/Config/Options#GridOptions">GridOptions</a> </div> ## Notes - **Fit-weighted composite.** Cells are Predict-style evaluations on attribute subsets × thresholds. The composite weights cells by fit so more reliable calibrations pull harder. - **Variable impact.** Impact on Fit and Impact on Prediction are task-specific total-importance measures (not only marginal effects). Prefer those names in analysis; technical aliases (MCTC / MCTP) are documented under [[Results/Grid Insights|Grid Insights]]. - **Retain is optional.** Insights above do not require retain. Set retain_grid_objects only when you need specific cell-level tables. - **Building blocks.** Fixed attributes and thresholds → [[Predict]]. Threshold search on one attribute set → [[MaxFit]]. Primary analysis → this page. ## Related - [[Functions/Overview|Functions]] - [[Predict]] · [[MaxFit]] - [[Config/Options#GridOptions|GridOptions]] · [[Results/Overview|Results]]