Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about CUR matrix decompositions and their applications in data analysis through this lecture by Michael Mahoney from Yahoo! Research. Explore how CUR decompositions provide an alternative to traditional matrix factorization methods like SVD by selecting actual rows and columns from the original matrix, making results more interpretable for data scientists and domain experts. Discover the theoretical foundations of CUR decompositions, including approximation guarantees and computational algorithms for efficiently computing these decompositions. Examine practical applications across various domains where maintaining interpretability of matrix factors is crucial, such as document analysis, gene expression data, and recommendation systems. Understand the trade-offs between approximation quality and interpretability when choosing between CUR and other matrix decomposition techniques. Gain insights into the algorithmic challenges and solutions for implementing CUR decompositions at scale, including randomized algorithms and sampling strategies for large-scale data matrices.
