network_stuff:machine_learning:supervised_learning
SUPERVISED LEARNING:
In python, we use the method fit to train a model. Fit~train
kmeans.fit(argument) kmeans.predict (argument) # Python predict() predicts the labels of the data values based on the trained model.
REGRESSION
For continuous tgt values
DECISON TREES:
- k-nearest neighbors
- ~ “classification by proximity” ; majority vote
- after doing it with all points, it creates a “boundary” (ie classification)
- Decision trees
- decision by path to leaves ; measure of center
- we ask question to narrow down areas (normally y/n Qs)
- decision trees can surface relationships that were not evident for the human understanding.
- Random forests: decision trees tend to overfitting. A solution is 'random forest'. Is a collection of decision trees (often hundreds of them), each trained differently on the same data,
network_stuff/machine_learning/supervised_learning.txt · Last modified: by 127.0.0.1