Survival prediction using microarray data – A topic modeling approach
Bibliographic record
Abstract
Many survival prediction models have been proposed over the years; some based on standard statistical survival analysis techniques, and others based on classic regression algorithms — eg, random survival forests, censored SVM. With the growing number of gene expression experiments being cataloged for analysis, we need to develop survival prediction models that can utilize such high dimensional data and also be interpretable. This study describes a way to learn survival prediction model that can accommodate such high dimensional data. We propose a novel approach based on topic modeling, called “discretized Latent Dirichlet Allocation (dLDA)”, that can derive interpretable and yet highly predictive covariates from the high dimensional microarray (gene expression) data. Latent Dirichlet Allocation is a widely used topic modeling generative model, with many successful applications in natural language (NL) processing. The LDA learner is an unsupervised technique that finds the inherent structure present in the data, by first finding topics, which are distributions over words (based on their integer counts in the document), and then decomposes each document into a Dirichlet mixture over these topics. This allows us to represent each document as this lower dimensional representation of the original text . We present a similar model that can be applied to our cohort of patients, each describes using real-valued gene expression data: here, we represents each patient [document] as a mixture over “(cancer) subtypes” [topics], where each subtype is a mixture over gene expression values [words]. Using the proposed dLDA as a preprocessing technique for the microarray data, we derive low-dimensional descriptions that can be used as input to a recent nonparametric learning algorithm, multi-task logistic regression (MTLR), that is designed to produce a model that can then predict patient-specific survival distributions — that is, given the (learned) MTLR model M, for each patient x, the personalized survival distribution p( t | M, x) can be inferred.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".