Recorded symptoms classify persistent smell and taste complaints after COVID-19
Source check: Source checked. Not independent scientific confirmation.
Finding: In a 413-person retrospective questionnaire dataset, learned models classified reported smell and taste complaints persisting beyond four weeks. The selected 83-record test reported AUCs up to 0.96 for taste and 0.97 for smell. This does not establish an early prospective forecast or improved treatment.
Technology: Random forest; XGBoost; LightGBM
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Fitted tree ensembles map questionnaire and symptom features to recorded persistent-complaint labels; the represented result is retrospective clinical-label evaluation.
- Test and result
- The study used a random 330-record training/validation set and 83-record test. Persistence predictors included symptom continuation after two weeks and smell/taste loss during infection.
- Limitation / still unknown
- The study is described as cross-sectional and does not establish that every predictor was recorded before outcome ascertainment. Whole-data feature ranking and test-informed algorithm selection limit generalization. Confusion-matrix paragraph denominators and percentages conflict with Table8; this record uses only table/ROC AUC ranges, not its accuracy or sensitivity claims. No external clinical validation, calibrated population risk or causal effect is established.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11336404/ · https://www.ncbi.nlm.nih.gov/research/bionlp/RESTful/pmcoa.cgi/BioC_json/PMC11336404/unicode
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Original primary clinical design, model training, persistence Results and Table8; archived publication metadata.
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 1/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.