Product Title Normalization using Phrase Indexer (E-Commerce)

<p>This use case demonstrates how <strong>M|Box’s Phrase Indexer</strong> and <strong>Approximate Phrase Index Matcher</strong> nodes can be combined to efficiently normalize product titles from e-commerce sources by matching them to a standardized internal product catalog.</p><p>Workflow Overview:</p><ol><li><p>The workflow loads:</p><ul><li><p>A reference table containing clean, standardized internal product names.</p></li><li><p>A second table containing e-commerce product titles, which may include additional attributes, reordered terms, or marketplace-specific wording</p></li></ul></li><li><p>The internal product names are indexed using the <strong>Phrase Indexer</strong>, creating a searchable phrase index optimized for approximate matching.</p></li><li><p>The e-commerce product titles are matched against this index using the <strong>Approximate Phrase Index Matcher</strong>. This allows robust matching even when:</p><ul><li><p>Titles contain additional descriptive terms (e.g., colors, storage sizes, marketing phrases)</p></li><li><p>Word order differs between title and reference name</p></li><li><p>Minor formatting variations occur</p></li></ul></li></ol><p>The result is a mapping from noisy marketplace titles to clean, normalized internal product names.</p>