Scribd
Upload
33 views

ML 2

Uploaded by

Utkarsh Bhalode
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
33 views

ML 2

Uploaded by

Utkarsh Bhalode
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 28

Supervised learning

Lecture – 2 (GNR 652)


Biplab Banerjee
Supervised learning
Supervised learning
Example - classification
Example - regression
Let’s look into features

 Categorical
 Ordinal
 Real-valued
 Integer-valued
Some examples
MNIST
Vector space at a glance
u,v – vectors, c,d - scalars
Subspace
Ideas of
 Linear independence
 Span
 Basis

We are interested about the Euclidean space


Metric space
Normed space

Any normed space is also a metric space!


Inner product space
Supervised learning
In a nutshell
Idea of concept
Subset of examples from the training data which uniquely
represent a given concept/class.
Hypothesis
 The function mapping which
 Maps (hopefully) correctly the inputs to the outputs
of the training set
 Generalizes well to previously unseen examples
 Approximates the true target function
Representation of the hypothesis
language
 Decision tree
 Linear regressor
 Multi-level perceptron
 Margin based techniques
 ...
Many hypothesis - Hypothesis
space
The task of concept learning
Idea of version space
What is a classifier then?
Inductive learning
Occum razor

Idea of inductive bias


Hoeffding inequality

Relation between the error/ risk in the training and test samples.
Both the errrors converge with increasing ‘m’

You might also like