Business Email Fuzzy Classification (Dictionary-Based)

<p>This workflow performs <strong>fuzzy classification</strong> of business emails using a <strong>dictionary-based approach</strong>.<br>It loads a dictionary of terms mapped to business constructs (e.g., <em>finance</em>, <em>hr_admin</em>), then uses fuzzy matching to find terms—even with typos—in each email’s subject and body. Matches are expanded, joined to constructs, and aggregated to produce <strong>transparent, explainable evidence</strong> per email and construct (counts, max similarity, distinct terms).<br>Use the final aggregates to assign labels (e.g., finance vs. hr_admin) and to power routing, analytics, or LLM grounding.</p>