In-Database Machine Learning for Big Data Entity Resolution

The data being collected by organizations is increasing relentlessly but it still can give a misleading or fragmented view of the real world. A person, for example, could appear multiple times or have multiple digital entities within the same database, due to typos, name changes, aggregation of different systems and so on. If we try to merge two databases, how do we match entities, when the ID systems might be different or contain errors? 

Entity resolution (ER) helps get to the truth. Entity resolution, which is the disambiguation of real-world entities in a database, is an essential data quality tool.

Graph provides an efficient approach for the entity resolution problem. A native graph database with massive parallel computing capability is the best tool to implement the approach.