Method and a system for fuzzy matching of entities in a database system based on machine learning

A non-transitory machine-readable storage medium that provides instructions that, if executed by a processor for estimating a number of distinct entities in a plurality of records of a database system, will cause said processor to perform operations comprising:

• receiving a first value of a first field type and a second value of the first field type;
• applying a first subset of a plurality of blurring operations on the first value and the second value to obtain a first set of blurred values;
• determining a first maximum score from first scores for blurred values of the first set of blurred values, wherein the first maximum score is greater than or equal to other scores of the first scores, and wherein each one of the first scores is indicative of a confidence that a match of the first value and the second value occurs with knowledge of a respective first blurred value from the first set of blurred values;
• determining a second maximum score from second scores for the blurred values of the first set of blurred values, wherein the second maximum score is greater than or equal to other scores from the second scores, and wherein each one of the second scores is indicative of a confidence that a non-match of the first value and the second value occurs with knowledge of the respective first blurred value from the first set of the blurred values;
• and
• responsive to determining that the first maximum score is greater than the second maximum score, outputting an indication that the first value matches the second value.