+ All Categories
Home > Documents > Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar...

Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar...

Date post: 15-Aug-2020
Category:
Upload: others
View: 1 times
Download: 0 times
Share this document with a friend
22
Margareta Ackerman Clustering
Transcript
Page 1: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Margareta Ackerman !

Clustering

Page 2: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Clustering is one of the most widely used tools for exploratory data analysis.

!

Its purpose is to identify similar groups of item in data.

What is clustering?

Page 3: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

!

Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city

Personalization: Identifying music you would like based on similar users Astronomy: Grouping starts based on their physical proximity

And much, much more! !

Clustering is applied virtually anywhere we find data.

!

Applications

Page 4: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Imagine points in Euclidean space. !

!

!

!

!

!

How would you divide them into k groups of similar items?

!

Come up with your own algorithm!

Page 5: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Clustering algorithms: A few classical examples

How can we partition data into k groups? !

!

!

!

Page 6: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Clustering algorithms: A few classical examples

How can we partition data into k groups? !

• Use Kruskal’s algorithm for MST (Single-linkage)

• Find the minimum cut (motivates spectral clustering methods)

• Find k “centers” that minimize the average distance to a center (k-median, k-means, ...)

• Many more... !

!

Page 7: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Clustering algorithms often have very different input-output behavior

Page 8: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

The “right solution" depends on the clustering application

This makes clustering inherently ambiguous.

Page 9: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

What will you gain from the clustering course?

• Explore the landscape of clustering algorithms.

• Study “The user’s dilemma,” which is the problem of selecting clustering algorithms for specific applications.

• Become an EXPERT in applying clustering in practice.

• Learn new research on how humans cluster.

• Tackle open problems in cluster theory and practice. !

!

!

!

!

!

Page 10: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

COMPUTATIONAL CREATIVITY

Page 11: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Computational creativity: Arthttp://www.aaronshome.com/aaron/gallery/index.html

Page 12: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Computational creativity: Arthttp://www.aaronshome.com/aaron/gallery/index.html

Page 13: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Computational creativity: Arthttp://www.aaronshome.com/aaron/gallery/index.html

Page 14: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

The Painting Foolhttp://www.thepaintingfool.com/

Hello there. I’m The Painting Fool. I’m a computer program that aspires

to be taken seriously as a creative artist in my own right - one day. I’ve

been programmed and trained by my teacher, Dr. Simon Colton, who is a senior lecturer here in the

Department of Computing. You can discover more about me at this website:

www.thepaintingfool.com

Page 15: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

The Painting Foolhttp://www.thepaintingfool.com/

Page 16: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

The Painting Foolhttp://www.thepaintingfool.com/

Page 17: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

The Painting Foolhttp://www.thepaintingfool.com/

Page 18: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Generative Music• https://aeigenfeldt.wordpress.com/

research/

Page 19: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Computational creativity: Food

• Cognitive cooking

• Cocktails

Page 20: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Creativity vs creativity• Small “c” creativity: Helping us be creative

• Casual creators (http://www.galaxykate.com/)

• To a lesser degree, all the systems discussed above

Page 21: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Creativity vs creativity• What is creativity?

• Can computers be creative?

• Can they surprise us?

• Can they be more creative than us?

Page 22: Clustering - brief introburmeste/Clustering_and... · Marketing: identifying target market (similar customers) Facility allocation: Placing hospitals in a new city Personalization:

Spring 2016: Computational creativity

• Get an overview of the field

• Create a casual creator

• Do a project on a topic of your choice

• If you are interested in taking this class, let me know soon!


Recommended