2 ms·
The optimization techniques required to actually fit models are almost all powered by some form of gradient descent, and integration is usually required in trul
by CuriouslyC 9y ago
The optimization techniques required to actually fit models are almost all powered by some form of gradient descent, and integration is usually required in truly probabilistic models to go from a density function to predictions.