Shawn Eary on 08 May 2020 21:34:31
Use Artificial Intelligence (AI) to lump several highly similar strings into categories.
As an example, suppose we have a field named Product that contains the following values: (Microphone; Desk; Microphone with Desk Stand; MicroPhonE; Mycrophone; Apple; Granny Smith Apple; Cannon Printer; CaNNNNon, 2020 Honda Odyssey [Free Vacation Package with Purchase]; Apple Tree ). I would then expect Power BI to be trainable so it could make the following categorizations: (Microphone -> Electronics; Desk -> Furniture; Microphone with Desk Stand -> Electronics; MicroPhonE -> Electronics; Mycrophone -> Electronics; Apple -> Grocery; Granny Smith Apple -> Grocery; Cannon Printer -> Electronics; CaNNNNon -> Weapons; 2020 Honda Odyssey [Free Vacation Package with Purchase] -> Automotive; Apple Tree -> Gardening )
As you can see above, the original field may contain typos and/or errors. By leveraging trainable AI, it may be possible to construct a new measure named Category from the original Product Field. That Category Measure would put the said products into “buckets” that could be used to easily analyze product data via the said categories.
For additional background see:
https://community.powerbi.com/t5/Desktop/Using-AI-to-Categorize-Titles-Subjects-and-Equipment-Names/m-p/1080897