Efficient subspace approximation algorithms
Nariankadu D. Shyamalkumar, Kasturi Varadarajan · Symposium on Discrete Algorithms · 2007
Confronted with high-dimensional data arising from either word-document count, global climate patterns or any one of the myriad other sources, most scientific approaches extract a good low-dimensional summary. This desire to reduce dimensionality may be seen as a consequence of Occam's Razor, and the scientific methodologies we have in mind include those from data mining and statistics.