Please enable JavaScript.
Coggle requires JavaScript to display documents.
Machine Learning - Coggle Diagram
Machine Learning
-
Supervised Learning
Explain this kind of learning
requires labels or desired output
for a given input. There are two
types of problems:
Regression and Classification
-
-
-
-
Parameter Estimation
MAP
Though this is not a full bayesian treatment
the addition of prior information (which is not uniform)
can be beneficial. This adds to regularization term
to the loss function.
MLE
Exlplain that how it leads to
least square error with certain
probability assumptions in regression
problems and leads to cross entropy
loss function in case of classification.
-
Unsupervised Learning
These algorithms do not require
any labels or there is no desired
output. These algorithms mainly
learn the structure of the data.
-
-
-
-