Boolean matrix factorization (BMF) is a well established and widely used tool for data analysis. Vast majority of existing algorithms for BMF is based on some greedy strategy which makes them highly sequential, thus unsuited for parallel execution. We propose a parallel variant of well-known BMF algorithm—GreConD, which is able to distribute workload among multiple parallel threads, hence can benefit from modern multicore CPUs. The proposed algorithm is based on formal concept analysis, intended for shared memory computers, and significantly reducing computation time of BMF via parallel execution.
CITATION STYLE
Krajča, P., & Trnecka, M. (2019). Parallelization of the GreConD algorithm for boolean matrix factorization. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11511 LNAI, pp. 208–222). Springer Verlag. https://doi.org/10.1007/978-3-030-21462-3_14
Mendeley helps you to discover research relevant for your work.