Projection Statistics – ProST: Online statistical assessment of group separation in data projection analysis
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Motivation Unsupervised data projection for the determination of trends in the data, visualization of multidimensional data in a reduced dimension space or feature space reduction through combination of data is a major step in data mining. Methods such as Principal Component Analysis or t-Distribution Stochastic Neighbor Embedding are regularly used as one of the first steps in computational biology or omics investigation. However, the significance of the separation of sample groups by these methods generally relies on visual assessment. User-friendly application for different projection methods, each focusing on distinct data properties, are needed as well as a rigorous method for statistical determination of the significance of separation of groups of interest in each dataset. Results We present Projection STatistics (ProST), a user-friendly solution for data projection analysis providing three unsupervised (PCA, t-SNE and UMAP) and one supervised (LDA) approach. For each method we are including a novel statistical investigation of the significance of group separation with Mann-Whitney U-rank or t-test analysis as well as necessary preprocessing steps. ProST provides an unbiased, objective application of the determination of the significance of the separation of measurement groups through either linear or manifold projection analysis with methods ranging from a focus on the separation of points based on major variances or on point proximities based on distance. Availability The ProST software application is freely available at https://complimet.ca/shiny/ProST/ with source code provided on https://github.com/complimet/prost . Contact danny.salem@nrc-cnrc.gc.ca or Miroslava.cuperlovic-culf@nrc-cnrc.gc.ca Supplementary information Supplementary help pages are provided at https://complimet.ca/shiny/ProST/ .
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.003 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it