55 Machine learning Clustering K means Clustering K means Feature 2 Source

# 55 machine learning clustering k means clustering k

• Test Prep
• 124

This preview shows page 66 - 74 out of 124 pages.

55

Subscribe to view the full document.

Machine learning – Clustering – K-means Clustering – K-means Feature #2 (Source: Naftali Harris) Feature #1 1 Consider data in R 2 spread on three different clusters, 2 Pick randomly K = 3 data points as cluster centroids, 3 Assign each data point to the class with closest centroid, 4 Update the centroids by taking the means within the clusters , 5 Go back to 3 until no more changes. 55
Machine learning – Clustering – K-means Clustering – K-means Feature #2 (Source: Naftali Harris) Feature #1 1 Consider data in R 2 spread on three different clusters, 2 Pick randomly K = 3 data points as cluster centroids, 3 Assign each data point to the class with closest centroid, 4 Update the centroids by taking the means within the clusters, 5 Go back to 3 until no more changes . 55

Subscribe to view the full document.

Machine learning – Clustering – K-means Clustering – K-means Optimal in terms of inter- and extra-class variability (loss), In practice, it requires much more iterations, Solutions strongly depend on the initialization , Good initializations can be obtained by K-means++ strategy. The number of class K is often unkown : usually found by trial and error, or by cross-validation, AIC, BIC, . . . K too small/large under/overfitting. (we will come back to this) The data dimension d is often much larger than 2 , subject to the curse of dimensionalty. (we will also come back to this) Vector quantization (VQ): the centroid substitutes all vectors of its class. 56
Machine learning – Classification Classification Classification: predict class d from observation x . (Source: Philip Martin) Classify a document into a predefined category. Documents can be text, images, videos. . . Popular ones are Support Vector Machines and Artificial Neural Networks. 57

Subscribe to view the full document.

Machine learning – Regression Regression Regression (prediction): predict value(s) from observation. Statistical process for estimating the relationships among variables. Regression means to predict the output value using training data. related to interpolation and extrapolation. Popular ones are linear least square and Artificial Neural Networks. 58
Machine learning – Classification vs Regression Classification vs Regression Classification Assign to a class Ex: a type of tumor is harmful or not Output is discrete/categorical v.s Regression Predict one or several output values Ex: what will be the house price? Output is a real number/continuous (Source: Ali Reza Kohani) Quiz, which one is which? denoising, identification, verification, approximation. 59

Subscribe to view the full document.

Machine learning – Polynomial curve fitting Polynomial curve fitting Consider N individuals answering a survey asking for their wealth: x i level of happiness: d i We want to learn how to predict d i (the desired output) from x i as d i y i = f ( x i ; θ ) where f is the predictor and y i denotes the predicted output.
• Fall '08
• staff

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern

Ask Expert Tutors You can ask 0 bonus questions You can ask 0 questions (0 expire soon) You can ask 0 questions (will expire )
Answers in as fast as 15 minutes