Methods and systems for biclustering algorithm

摘要

Methods and systems for improved unsupervised learning are described. The unsupervised learning can consist of biclustering a data set, e.g., by biclustering subsets of the entire data set. In an example, the biclustering does not include feeding know and proven results into the biclustering methodology or system. A hierarchical approach can be used that feeds proven clusters back into the biclustering methodology or system as the input. Data that does not cluster may be discarded. Thus, a very large unknown data set can be acted on to learn about the data. The system is also amenable to parallelization.

主权项

1. A method for clustering information from a data set, comprising:
creating first clusters of related data from a first subspace of unclustered data in the data set, wherein related data added to each one of the first clusters shares at least one attribute in the first subspace; and creating second clusters of related data from a second subspace of unclustered data in the data set, wherein adding the related data from the second subspace to one of the second clusters requires satisfying a vigilance parameter to ensure each one of the second clusters of related data shares at least one attribute in the second subspace; and providing feedback to increase the vigilance parameter if the related data from the second subspace added to the one of the second clusters does not share at least one attribute associated with one of the first clusters, wherein increasing the vigilance parameter forces the related data from the second subspace to be added to a new or different one of the second clusters.