2 ms·
Agreed. For many prediction / classification tasks you don't need much feature engineering. Or much data for that matter ;)
by fitzwatermellow 10y ago
Agreed. For many prediction / classification tasks you don't need much feature engineering. Or much data for that matter ;)
- antirez 10y agoYep... nor there is locality in the inputs to exploit like it happens in CNNs.