I am trying to map some categorical inputs combined with customer unique id to a single result property. So my input will be a numeric (discrete) id and an array indicates for each property weather it contains the value or not. For example, consider I have 5 types of item where in a specific instance item one will output item 5, I’d like my data row to look like:
[3326,[1,0,0,0,0]] -> [0,0,0,0,1]
However I couldn’t find a way to model this with layers.
If I using a different approach where a row is converted into a single array such as [3326, 1] -> 5 The results are not good.
Anyone?
question from:https://stackoverflow.com/questions/65871052/tensorflowjs-layer-of-both-data-and-identifier
