Real-Time Entity Resolution
Understand use cases of real-time matching and how it fits into your entity resolution framework.
We'll cover the following...
A website matches visitors using cookies and geocoded IPs to provide personalized content. A call center agent needs a quick summary of a customer’s interactions across all channels. A payment service must decide about fraud or no fraud within seconds for an endless stream of incoming requests. All these use cases require insights at query time, whereas most we have discussed before function with precalculated entity resolution.
That does not mean that query-time response is without precalculation. It is like with machine learning—train the model on a large batch of data (precalculate), deploy the model as a web service, and that responds with predictions to small batches of new inputs (query-time).
Prevent duplicates with autocomplete
Google Maps autocompletes the user’s search to select one of a few best matches. The user also learns how to change the query interactively from the instant feedback. Similarly, we can benefit from autocomplete at every interface where our users create new records.
With autocomplete, ...