mi@0: #!/usr/bin/python
mi@0: #
mi@0: # Copyright (C) Christian Thurau, 2010. 
mi@0: # Licensed under the GNU General Public License (GPL). 
mi@0: # http://www.gnu.org/licenses/gpl.txt
mi@0: """  
mi@0: PyMF Simplex Volume Maximization for CUR [1]
mi@0: 
mi@0:     SIVMCUR: class for SiVM-CUR
mi@0: 
mi@0: [1] C. Thurau, K. Kersting, and C. Bauckhage. Yes We Can - Simplex Volume 
mi@0: Maximization for Descriptive Web-Scale Matrix Factorization. In Proc. Int. 
mi@0: Conf. on Information and Knowledge Management. ACM. 2010.
mi@0: """
mi@0: 
mi@0: 
mi@0: import numpy as np
mi@0: import scipy
mi@0: from sivm import SIVM
mi@0: from cur import CUR
mi@0: 
mi@0: __all__ = ["SIVM_CUR"]
mi@0: 
mi@0: class SIVM_CUR(CUR):
mi@0:     '''
mi@0:     SIVM_CUR(data, num_bases=4, dist_measure='l2')
mi@0:     
mi@0:     Simplex Volume based CUR Decomposition. Factorize a data matrix into three 
mi@0:     matrices s.t. F = | data - USV| is minimal. Unlike CUR, SIVMCUR selects the
mi@0:     rows and columns using SIVM, i.e. it tries to maximize the volume of the
mi@0:     enclosed simplex.
mi@0:     
mi@0:     Parameters
mi@0:     ----------
mi@0:     data : array_like [data_dimension x num_samples]
mi@0:         the input data
mi@0:     rrank: int, optional 
mi@0:         Number of rows to sample from data.
mi@0:         4 (default)crank
mi@0:     crank: int, optional
mi@0:         Number of columns to sample from data.
mi@0:         4 (default)   
mi@0:     dist_measure: string, optional
mi@0:         The distance measure for finding the next best candidate that 
mi@0:         maximizes the simplex volume ['l2','l1','cosine','sparse_graph_l2']
mi@0:         'l2' (default)
mi@0:     
mi@0:     Attributes
mi@0:     ----------
mi@0:         U,S,V : submatrices s.t. data = USV        
mi@0:     
mi@0:     Example
mi@0:     -------
mi@0:     >>> import numpy as np
mi@0:     >>> data = np.array([[1.0, 0.0, 2.0], [0.0, 1.0, 1.0]])
mi@0:     >>> sivmcur_mdl = SIVM_CUR(data, show_progress=False, rrank=1, crank=2)    
mi@0:     >>> sivmcur_mdl.factorize()
mi@0:     '''
mi@0:     
mi@0:     def __init__(self, data, k=-1, rrank=0, crank=0, dist_measure='l2', init='origin'):
mi@0:         CUR.__init__(self, data, k=k, rrank=rrank, crank=rrank)
mi@0:         self._dist_measure = dist_measure
mi@0:         self.init = init
mi@0: 
mi@0:     def sample(self, A, c):
mi@0:         # for optimizing the volume of the submatrix, set init to 'origin' (otherwise the volume of
mi@0:         # the ordinary simplex would be optimized) 
mi@0:         sivm_mdl = SIVM(A, num_bases=c, dist_measure=self._dist_measure,  
mi@0:                             init=self.init)                        
mi@0:         sivm_mdl.factorize(show_progress=False, compute_w=True, niter=1,
mi@0:                            compute_h=False, compute_err=False)
mi@0:         
mi@0:         return sivm_mdl.select    
mi@0:             
mi@0:             
mi@0:     def factorize(self):
mi@0:         """ Factorize s.t. CUR = data
mi@0:             
mi@0:             Updated Values
mi@0:             --------------
mi@0:             .C : updated values for C.
mi@0:             .U : updated values for U.
mi@0:             .R : updated values for R.          
mi@0:         """            
mi@0:         # sample row and column indices that maximize the volume of the submatrix
mi@0:         self._rid = self.sample(self.data.transpose(), self._rrank)
mi@0:         self._cid = self.sample(self.data, self._crank)
mi@0:                         
mi@0:         self._rcnt = np.ones(len(self._rid))
mi@0:         self._ccnt = np.ones(len(self._cid))    
mi@0:                                     
mi@0:         self.computeUCR()
mi@0: 
mi@0: 
mi@0: if __name__ == "__main__":
mi@0:     import doctest  
mi@0:     doctest.testmod()