I write this topic to ask your suggestion for a[b] Reference Data[/b] project.Problem: To make a match of transactional products with catalog reference products and learn about the matches already made. This match should be made by a set of attributes (similarity of name, similarity of prices, etc.) In practice I get daily transactions with a code and a description and I have to match with my catalog reference products. As the codes are not the same, I have to make this match by description. However, the description can be changed and may not be exactly the same.I plan to create a model that will learn from matches already made: I want to create a process of association / trust (with the products that matched successfuly) to determine not only the degree of similarity, but ensure association between them. For this purpose I was thinking about using:- MS Azure Machine Learning- MS Data Quality Services- ...Do you guys have any suggestions to solve problems of this kind? What is for you the best solution?
↧