K-Means clustering Classes英文版本.pdf

Clustering with the K-Means Algorithm Andrew Ng (video tutorial from Coursera’s “Machine Learning” class) Transcript written by José Soares Augusto, June 2012 (V1.0) 1 Unsupervised Learning_ Introduction I’d like to start to talk about clustering, our first unsupervised learning algorithm, where we learn from unlabeled data instead of from labeled data. I briefly talked about unsupervised learning at the beginning of the class, but now it’s useful to contrast it with supervised learning. In a typical supervised learning problem we are given a labeled training set, and the goal is to find the decision boundary that separates the positive and the negative label examples, and to fit a hypothesis to it (blue line). In the unsupervised learning problem we’re given data that does not have any labels associated with it. And so our training set is written just x 1, x 2 and so on, up to x m, and we don’t get any labels y . So, in unsupervised learning we give this sort of unlabeled training set to an algorithm, and we just ask the algorithm: find some structure in the data for us. One type of structure we might have an algorithm find is shown in the data set above, which has points grouped into two separate clusters, and so an algorithm that finds those clusters (green circles) is called a cluster- ing algorithm. This will be our first type of unsupervised learning, although there are other types of unsupervised learning algorithms that find other types of structure, or other types of patterns in the data, other than clusters. With help from the English subtitles which are available in the “Machine Learning” site, /ml/lecture/. 1 So what is clustering good for? One application is market segmentation, where you may have a database of customers a