PMC:4996402 / 41150-42375 JSONTXT

{"target":"https://pubannotation.org/docs/sourcedb/PMC/sourceid/4996402","sourcedb":"PMC","sourceid":"4996402","source_url":"https://www.ncbi.nlm.nih.gov/pmc/4996402","text":"7. Conclusions\nThis paper surveys the state-of-the-art NCA-based algorithms proposed in the literature. These algorithms rely on a linear model and concentrate on reconstructing the TFA matrix and the connectivity matrix by using the information provided by microarray gene expression data. The algorithms reviewed herein can be divided broadly into two categories: iterative and non-iterative methods. For the iterative methods, the estimation process for the connectivity matrix and TFA matrix starts with an initial guess, and then, it proceeds through a sequence of iterative steps. The output of each step is fed as an input to the next step. On the other hand, the non-iterative methods aim to overcome the drawbacks of iterative methods, especially to reduce the high computational complexity by reformulating the NCA problem. A summary of the surveyed NCA-based algorithms is illustrated in Table 2 for further details.\nmicroarrays-04-00596-t002_Table 2 Table 2 Summary of NCA-based algorithms. mNCA, motif-directed NCA; gNCA, generalized NCA; NCAr, revised NCA; gfNCA, generalized-framework NCA; nnNCA, non-negative NCA; ALS, alternate least-squares; SSP, subspace separation principle; TLS, total least-squares. ","divisions":[{"label":"Title","span":{"begin":0,"end":14}},{"label":"Table caption","span":{"begin":928,"end":1224}}],"tracks":[]}