Mathematics, Volume 12, Issue 21
2024 November-1 - 142 articles
Cover Story: High-dimensional sparse matrix data, like word-word co-occurrence counts and e-commerce item pairs, often have skewed non-negative values with many zeros. We predict item or user relevance using unknown dense vector representations. Our model employs zero-inflated Gamma random variables and cosine similarities. We estimate the unknown parameters with shared parameter alternating zero-inflated Gamma regression models (SA-ZIGs), considering canonical and log link models. We propose two parameter updating schemes and an algorithm for estimating parameters, as well as an analytical convergence analysis. Numerical studies show that a SA-ZIG with a learning rate adjustment performs well, while Fisher scoring without adjustment may fail to find the maximum likelihood estimate. View this paper - Issues are regarded as officially published after their release is announced to the table of contents alert mailing list .
- You may sign up for email alerts to receive table of contents of newly released issues.
- PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.