A reformulation of pLSA for uncertainty estimation and hypothesis testing in bio-imaging

Tar, Paul; Thacker, Neil A.; Deepaisarn, Somrudee; O’Connor, James P.B.; McMahon, Adam

doi:10.1093/bioinformatics/btaa270

Cited by 3 publications

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To avoid a zero probability of unseen documents in the training phase, Brants has introduced P(d i ) > 0, stating that the log-likelihood can be maximized, taking into account only the second term of (13), and for the new documents likelihood L is…”

Section: Probabilities For Unseen Documentsmentioning

confidence: 99%

“…This section also explains its use as a semi-supervised technique, achieved by modifying the likelihood function to assign probabilities to unobserved documents [9]; the hypothesis of the case of Boolean variables [10], important in bioinformatics, and the hypothesis to extend the model to continuous data classes [11,12]. This multivariate model lets inferential applications [13]. To complete this section, we describe several fields of applications in Section 4.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Revisiting Probabilistic Latent Semantic Analysis: Extensions, Challenges and Insights

Figuera,

García Bringas

2024

Technologies

View full text Add to dashboard Cite

This manuscript provides a comprehensive exploration of Probabilistic latent semantic analysis (PLSA), highlighting its strengths, drawbacks, and challenges. The PLSA, originally a tool for information retrieval, provides a probabilistic sense for a table of co-occurrences as a mixture of multinomial distributions spanned over a latent class variable and adjusted with the expectation–maximization algorithm. The distributional assumptions and the iterative nature lead to a rigid model, dividing enthusiasts and detractors. Those drawbacks have led to several reformulations: the extension of the method to normal data distributions and a non-parametric formulation obtained with the help of Non-negative matrix factorization (NMF) techniques. Furthermore, the combination of theoretical studies and programming techniques alleviates the computational problem, thus making the potential of the method explicit: its relation with the Singular value decomposition (SVD), which means that PLSA can be used to satisfactorily support other techniques, such as the construction of Fisher kernels, the probabilistic interpretation of Principal component analysis (PCA), Transfer learning (TL), and the training of neural networks, among others. We also present open questions as a practical and theoretical research window.

show abstract

Section: Probabilities For Unseen Documentsmentioning

confidence: 99%