3 ms·
Agree with your point about anonymization, but here is a call out for generated data too: if the objective is to build a model, you might end up losing informa
by abhgh 6y ago
Agree with your point about anonymization, but here is a call out for generated data too: if the objective is to build a model, you might end up losing information, or worse, your model might end up modelling the assumptions in the generation process.
- jacquesm 6y agoYes, that's absolutely valid. This too is hard. But then again, if it were easy everybody would be doing it so consider it a problem that when solved properly becomes part of your moat and a thing that you could easily mention in a sales process to ensure a level playing field with other parties pitching.