| 摘要 |
Systems and methods that enhance estimate(s) of features (e.g., word associations), via employing a sampling component (e.g., sketches) that facilitates computations of sample contingency tables, and designates occurrences (or absence) of features in data (e.g., words in document lists). The sampling component can further include a contingency table generator and an estimation that employs a likelihood argument (e.g., partial likelihood, maximum likelihood, and the like) to estimate features/word pair(s) associations in the contingency tables.
|
