<div class="section-label"><span class="num">01</span> Functions</div>
<div class="section-title">Start with <em>grid</em> prediction.</div>
Relevance-based prediction has three ideas that work together: **relevance** (which past cases matter), **fit** (how reliable this task looks before you act), and **grid prediction** (how to blend many calibrations toward the ones that fit best).
For day-to-day analysis, use [[Grid Prediction]]. It is the form of RBP we want teams to adopt. [[Predict|Predict]] (partial-sample regression) and [[MaxFit]] remain available. They are important building blocks with clear uses, but they are not the default path.
Install and licensing stay under [[Install/Get Started|Install]]. Settings live under [[Config/Overview|Config]]. What you get back lives under [[Results/Overview|Results]].
<div class="pillars">
<div class="pillar">
<div class="pillar-num">↳ 01</div>
<div class="pillar-title">Grid Prediction</div>
<div class="pillar-text">Blend many attribute × observation calibrations by fit into one composite, with variable impact measures.</div>
</div>
<div class="pillar">
<div class="pillar-num">↳ 02</div>
<div class="pillar-title">Building blocks</div>
<div class="pillar-text">Predict and MaxFit power cells and threshold search. Use them when you need a fixed setup or a focused diagnostic.</div>
</div>
<div class="pillar">
<div class="pillar-num">↳ 03</div>
<div class="pillar-title">Relevance scores</div>
<div class="pillar-text">Inspect relevance, similarity, and informativeness on their own when you are not ready to forecast yet.</div>
</div>
</div>
<p> </p>
<div class="section-label"><span class="num">02</span> Choose a path</div>
<div class="solution-list">
<div class="solution-row">
<div class="solution-num">/01</div>
<div class="solution-title"><a href="/Functions/Grid%20Prediction">Grid Prediction</a></div>
<div class="solution-desc">Primary analysis path. Diversify across attribute combinations and relevance thresholds, then form one fit-weighted composite with Impact on Fit and Impact on Prediction.</div>
</div>
<div class="solution-row">
<div class="solution-num">/02</div>
<div class="solution-title"><a href="/Functions/Predict">Predict</a></div>
<div class="solution-desc">Partial-sample regression at thresholds you choose. It is the evaluation inside each grid cell; use it directly when the attribute set and cuts are already fixed.</div>
</div>
<div class="solution-row">
<div class="solution-num">/03</div>
<div class="solution-title"><a href="/Functions/MaxFit">MaxFit</a></div>
<div class="solution-desc">Search thresholds (and optionally both censor types) for the strongest fit on a fixed attribute set. Useful diagnostics; not a substitute for Grid.</div>
</div>
<div class="solution-row">
<div class="solution-num">/04</div>
<div class="solution-title"><a href="/Functions/Relevance%20Metrics">Relevance Metrics</a></div>
<div class="solution-desc">Score relevance, similarity, and informativeness without forming a full prediction.</div>
</div>
</div>
All prediction calls take the same inputs: outcomes <span class="mono">y</span>, attributes <span class="mono">X</span>, and circumstances <span class="mono">theta</span>. Leave options out to use sensible defaults.
> [!tip]
> New to the engine? Call [[Grid Prediction|predict_grid]] first. Reach for [[Predict]] or [[MaxFit]] when you already know you need a single attribute set or a threshold-only search.
<div class="btn-row center">
<a class="btn-primary" href="/Functions/Grid%20Prediction">Start with Grid Prediction</a>
<a class="btn-ghost" href="/Results/Grid%20Insights">See Grid insights</a>
</div>