Presentation is loading. Please wait.

Presentation is loading. Please wait.

Introduction to Classification & Clustering Villanova University Machine Learning Lab Module 4.

Similar presentations


Presentation on theme: "Introduction to Classification & Clustering Villanova University Machine Learning Lab Module 4."— Presentation transcript:

1 Introduction to Classification & Clustering Villanova University Machine Learning Lab Module 4

2 Machine Learning Getting a computer to learn from data A type of Artificial Intelligence, where a computer does something "intelligent" Ability of a computer to improve what it does in a way that mimics how humans learn, like with repetition or experience

3 Examples of Machine Learning Face detection used by Facebook to help you automatically tag friends Spam filters that get better over time at identifying and trashing spam emails Fraud detection that notices suspicious patterns of credit card use and you get a call Optical character recognition that reads the numbers written on a check you deposit

4 Classification & Clustering Two ways to put things into categories Classification – Categories already exist – Put each thing into the category where it fits best Clustering – Categories don’t yet exist – Put the things into brand new categories based on similar characteristics or features of the things

5 Classification Some of the things being put into groups are already labeled with their "class." These class labels are used to guide or supervise the classification of unlabeled things into one of the classes. When the classes are known ahead of time this type of machine learning is called: Supervised Learning

6 Example: Classification DogsCats Classify the photos on the next slide into one of these two categories

7 Example: Classification

8 How Did Classification Work? How did you approach the task? What made it easy to do? What made it hard to do? How good are your final categories?

9 Clustering None of the things being put into groups already have their class labels so the grouping is unsupervised. The task becomes figuring out clusters of things with similar features. When the classes are unknown ahead of time this type of machine learning is called: Unsupervised Learning

10 Example: Clustering Cluster the photos on the next slide into two categories. You decide! Group 2Group 1

11 Example: Clustering

12 How Did Clustering Work? How did you approach the task? What made it easy to do? What made it hard to do? How does it compare with Classification? How good are your final categories?

13 Example: Clustering Again! Cluster the photos on the next slide into three categories. You decide! Group 2Group 1Group 3

14 Example: Clustering

15 How Did Clustering Work? How did you create categories? Did it get harder to do with more categories? How good are your results?


Download ppt "Introduction to Classification & Clustering Villanova University Machine Learning Lab Module 4."

Similar presentations


Ads by Google