Best approach to use to match new data against already existing indexed data


I have an items database with an elastic search index used for simple matching. I would like to leverage elastic search so that before new items are added, it can be determined if an existing, similar item already exists. An item has a name. New items being added generally have much longer and more descriptive names than existing items. Once a new item is matched to an existing item, I would like the new item's description to be included in attempts to match future items.

(Doug Turnbull) #2

Have you looked into more like this?

Do you just want to put this on autopilot? Or is ther a human in the loop? On autopilot, more like this could be risky. There's always a chance that the candidate document is not the same with minor edits. However you CAN tune more like this to be very strict (see docs above).


Thanks for the pointer Doug! No, I have not seen the more like this
query. It looks promising. Yes, there will be a human in the loop to
double check the mapping and select from prospective candidates. However,
where the relevance is high, I am hoping we can get to an auto pilot

(system) #4