You work as a machine learning specialist for a retail chain that has recently purchased another retail chain and is in the process of merging the two chain’s systems. Both retail chains have customer databases. Some of the firm’s customers overlap, meaning that the same customer registered with both chains in the past. When merging the customer data stores of the two, presently merged retail chains, you need to link duplicate customer data to have one accurate customer data source.
You have been assigned to create the new customer data source for the presently merged retail chain. Instead of trying to find duplicate customer data manually through traditional programming techniques, you have decided to use machine learning techniques to solve the problem.
You have determined that the AWS Glue Machine Learning FindMatches Transform is the best solution to this problem. Knowing that incorrectly linking what appear to be duplicate customers must be avoided at all costs, how should you configure the AWS Glue FindMatches ML Transform parameters to achieve the most efficient and accurate duplicate customer detection process?

Answer options:

A.Set the FindMatches precision-recall parameter to ‘precision’ and the accuracy-cost parameter to ‘accuracy’.
B.Set the FindMatches precision-recall parameter to ‘precision’ and the accuracy-cost parameter to ‘lower cost’.
C.Set the FindMatches precision-recall parameter to ‘recall’ and the accuracy-cost parameter to ‘accuracy’.
D.Set the FindMatches precision-recall parameter to ‘recall’ and the accuracy-cost parameter to ‘lower cost’.

Answer: A
Option A is correct. Setting the FindMatches precision-recall parameter to ‘precision’ minimizes false positives (when you don’t have a match of a duplicate customer but mark it as a match mistakenly). This is what you want. Setting the FindMatches accuracy-cost parameter to ‘accuracy’ maximizes the transform accuracy of finding matching records as duplicate. This is also what you want.
Option B is incorrect. Setting the FindMatches precision-recall parameter to ‘precision’ minimizes false positives (when you don’t have a match of a duplicate customer but mark it as a match mistakenly). This is what you want. But, setting the accuracy-cost parameter to ‘lower cost’ favors cost or the speed of running the transform at the expense of the transform’s accuracy. This may make your transform more performant, but your primary concern is avoiding linking customers incorrectly. So you should set the accuracy-cost parameter to ‘accuracy’.
Option C is incorrect. Setting the FindMatches precision-recall parameter to ‘recall’ minimizes false negatives (when you have a match of a duplicate customer but fail to detect it). This may cause customer frustration, but your primary concern is avoiding linking customers incorrectly.
Option D is incorrect. Setting the FindMatches precision-recall parameter to ‘recall’ minimizes false negatives (when you have a match of a duplicate customer but fail to detect it). This may cause customer frustration, but your primary concern is avoiding linking customers incorrectly.
Reference:
Please see the AWS Glue developer guide titled Machine Learning Transforms in AWS Glue, and the AWS Glue developer guide titled Tuning Machine Learning Transforms in AWS Glue.