Noisy and Incomplete Boolean Matrix Factorizationvia Expectation Maximization
This work addresses a domain-specific problem in data analysis for applications requiring robust factorization, but it is incremental as it builds on existing probabilistic approaches.
The study tackled the problem of Boolean matrix factorization in the presence of noise and missing values by proposing a new probabilistic algorithm that eliminates assumptions about latent factors, and it showed favorable performance compared to state-of-the-art probabilistic algorithms in real data experiments.
Probabilistic approach to Boolean matrix factorization can provide solutions robustagainst noise and missing values with linear computational complexity. However,the assumption about latent factors can be problematic in real world applications.This study proposed a new probabilistic algorithm free of assumptions of latentfactors, while retaining the advantages of previous algorithms. Real data experimentshowed that our algorithm was favourably compared with current state-of-the-artprobabilistic algorithms.