Comparing RoBiC with other Systems

• SVD based approaches

• Sven Bergmann, Jan Ihmels, and Naama Barkai, Iterative signature algorithm for the analysis of large-scale gene expression data, Physical Review E, 67, 031902 (2003).
  Similarities:
  • both use something like SVD, with a threshold
  • both work iteratively
  • both finds biclusters that can overlap
  • both extract biclusters based on the "larger" values (of something...)
  Differences:
  • RoBiC considers values in eigenvector, not in matrix itself
  • RoBiC does not require the user to specify a threshold
  • ISA requires the data matrix to be normalized (twice)... RoBiC does not require this process
  • RoBiC (well, BiC) explicitly considers the task of producing a classifier