no matches found in two tables even though they are present - AWS Entity Resolution

Hi,

In the given example, there may not be enough context provided to the Machine Learning system to produce matches as expected. This is because only a 'Name' as input may not be sufficient (as there could be multiple individuals with the same name). So, to improve this and get results as expected, you may consider adding another input such as a 'Phone Number'. This will then give more context to the Machine Learning system which will then provide more confident results. To test this, I used the same data as you and provided a 'Phone Number' field to both datasets and was then able to get matches as expected (even for names such as 'leslie kingkong').

Alternatively, if you want to keep using the same tables as they are, a rule-base matching could be more suitable here. I have tested this from my end and it is also working as expected.

If you are still facing issues despite the above recommendations, I would suggest raising a case with AWS Support who would be able to assist further. You can do so here.

Hope this helps!

Hello, thank you for clarifying the ML-based match for me - I can see now that it won't be the right choice in our case - we only have one column to match with - we want to find similar names, in case they are mispronounced or in a different language (and find the same in another language). I've tried the rule-based matching like you suggested - matching only by "author_name" and "hcp_names" and unfortunately I also didn't get the desired results - each time I'm running this way I'm getting column named "Error.dynamic Recod.match Rule" with values "NoRule" in each row. I'm not sure why this error as output though.