Driving with Data#
Recall:
- Rules-based Methods
Prior-knowledge is encoded explicitly as rules over patterns.
Behavior is acheived implicitly through application of the rules to matched patterns.
So:
- Data-Driven Methods:
Desired Behavior is provided explicitly as observations to be reproduced or mimiced
Rules and domain-knowledge is learned implicitly via a model, through statistical patterns or minimization of error
Tip
This Chapter’s extra reading:
Pomegranate library documentation, and examples for Markov Models, HMMs.