PALTiling
swMATH ID:  28419 
Software Authors:  Hess, Sibylle; Morik, Katharina; Piatkowski, Nico 
Description:  The PRIMPING routine – Tiling through proximal alternating linearized minimization. .. We introduce PALTiling, a general framework to compute tilings according to a cost measure based on a theoretically founded numerical optimization technique. Requiring that the cost measure has a smooth relaxed function, which combines the matrix factorization error with a regularizing function, PALTiling minimizes the relaxed objective under convergence guarantees. To simulate the minimization subject to the constraint that the matrices are binary, we derive a closed form of the proximal mapping with respect to a function which penalizes nonbinary values. A thresholding to binary values according to the actual cost measure enables an automatic determination of the factorization rank. 
Homepage:  https://link.springer.com/article/10.1007/s106180170508z 
Keywords:  tiling; Boolean matrix factorization; minimum description length principle; proximal alternating linearized minimization; nonconvexnonsmooth minimization; alternating minimization 
Related Software:  PROXIMUS; ComiRNet; FABIA; PaNDa+; Krimp; StreamKrimp; MDL4BMF 
Cited in:  2 Documents 
Standard Articles
1 Publication describing the Software, including 1 Publication in zbMATH  Year 

The PRIMPING routine – tiling through proximal alternating linearized minimization. Zbl 1409.68233 Hess, Sibylle; Morik, Katharina; Piatkowski, Nico 
2017

all
top 5
Cited by 6 Authors
2  Hess, Sibylle 
1  Ceci, Michelangelo 
1  Hochstenbach, Michiel E. 
1  Morik, Katharina 
1  Piatkowski, Nico 
1  Pio, Gianvito 
Cited in 1 Serial
2  Data Mining and Knowledge Discovery 
Cited in 4 Fields
1  Linear and multilinear algebra; matrix theory (15XX) 
1  Statistics (62XX) 
1  Numerical analysis (65XX) 
1  Computer science (68XX) 