Scribd
Upload
2 views9 pages

CE802_Lec_IBL_add_slides

The document discusses instance-based learning, focusing on assumptions, inductive bias, and similarity functions. It emphasizes the importance of encoding and distance measures in classification tasks, particularly in the context of nearest neighbor algorithms (1-NN and k-NN). The role of the parameter 'k' in determining the effectiveness of the model is also highlighted.

Uploaded by

Anand A J
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
2 views9 pages

CE802_Lec_IBL_add_slides

The document discusses instance-based learning, focusing on assumptions, inductive bias, and similarity functions. It emphasizes the importance of encoding and distance measures in classification tasks, particularly in the context of nearest neighbor algorithms (1-NN and k-NN). The role of the parameter 'k' in determining the effectiveness of the model is also highlighted.

Uploaded by

Anand A J
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/ 9

Instance-based learning Assumptions and inductive bias

Smooth relative to the sampling of feature space


x2

x1

CE802 (CSEE) Instance-based learning 5 / 24


Instance-based learning Assumptions and inductive bias

Smooth relative to the sampling of feature space


x2

x1

CE802 (CSEE) Instance-based learning 5 / 24


Instance-based learning Assumptions and inductive bias

Smooth relative to the sampling of feature space


x2

x1

CE802 (CSEE) Instance-based learning 5 / 24


Instance-based learning Similarity functions

Choosing the right encoding/distance

Imagine we are designing autopilot able to fly a broken plane.


We use a simple encoding:
1 binary attribute: “left engine broken”
1 binary attribute: “right engine broken”
and a simple distance: number of attributes that are different.

CE802 (CSEE) Instance-based learning 10 / 24


Instance-based learning Classification functions

Building 1-nn
x2

x1

CE802 (CSEE) Instance-based learning 14 / 24


Instance-based learning Classification functions

Building k-nn
x2

x1

CE802 (CSEE) Instance-based learning 15 / 24


Instance-based learning Classification functions

Role of k: training set

CE802 (CSEE) Instance-based learning 16 / 24


Instance-based learning Classification functions

Role of k: 1-NN

CE802 (CSEE) Instance-based learning 16 / 24


Instance-based learning Classification functions

Role of k: 3-NN

CE802 (CSEE) Instance-based learning 16 / 24

You might also like