• Unsupervised learning: There are neither labeled data for training nor any rewards from the environment. What would the machine learn in this case? The goal is to make the machine learn the patterns of data and be useful for future predictions. Classic examples of unsupervised learning are clustering and dimensionality reduction. Common clustering algorithms include k-means (based on centroid models), a mixture of Gaussians, hierarchical clustering (based on connectivity models), and an expectation maximization algorithm (which uses a multivariate normal distribution model). The various dimensionality reduction techniques include the factor analysis, principal component analysis (PCA), independent component analysis (...
- 3. Realizing Machine Learning Algorithms with Spark
- from Big Data Analytics Beyond Hadoop: Real-Time Applications with Storm, Spark, and More Hadoop Alternatives
- Publisher: Pearson Business
- Released: May 2014
use this for next week's rollout
Share this highlighthttp://www.safaribooksonline.com/a/big-data-analytics/78141/