Your memory might be more fallible than you think: from only 0.25% of early SARS-CoV-2 infections being detected in the first symptom week to recall of key health details ranging as low as 22% false recall and leaving 45% of exposed contacts infected among those tested, this post explores how often recall succeeds, fails, and reshapes health estimates.
Key Takeaways
- 0.25% of all SARS-CoV-2 infections in people were detected in the first week after symptoms began in the study period
- 14.7% of participants without symptoms were PCR positive
- 44.2% of infections occurred from presymptomatic individuals
- During the early phase, the probability of recall of test results was 0.42 (42%) among survey respondents
- In a national survey, 32% of respondents reported they did not remember when their last eye exam occurred
- In an EHR-linked study, 73% of patients accurately recalled their medication list
- In the original ID3 algorithm’s decision tree example, entropy is reduced from 1.0 to 0.0 after splitting on the attribute with information gain 1.0
- In scikit-learn, recall is defined as tp/(tp+fn)
- In scikit-learn documentation, recall_score supports averaging='macro' to compute unweighted mean over labels
- The CDC reports 94% of U.S. adults reported being in contact with a doctor at least once in the past year (health care access survey)
- The US USPSTF recommends breast cancer screening: 2024 draft recommendation for women aged 40-74 (screening interval 2 years)
- USPSTF recommends colorectal cancer screening for adults 45-75, with annual FIT or colonoscopy intervals (1 year for FIT)
- In the “TREC Precision-Recall” experiments, recall is plotted on x-axis from 0 to 1
- In the standard IR definition, recall = TP/(TP+FN) equals sensitivity for retrieval contexts
- The Recall metric in recommendation systems is “fraction of relevant items retrieved”; definition is stated in RecBole docs
Recall is unreliable: infections, tests, vaccines, and meds often forgotten or misremembered.
Case & Detection Rates
10.25% of all SARS-CoV-2 infections in people were detected in the first week after symptoms began in the study period[1]
Verified214.7% of participants without symptoms were PCR positive[2]
Verified344.2% of infections occurred from presymptomatic individuals[3]
Verified459% of transmissions were from presymptomatic or asymptomatic individuals[4]
Directional555% of people who were tested after exposure had not been infected, implying a 45% infection rate among tested exposed contacts[5]
Single sourceCase & Detection Rates Interpretation
In this study, the virus was already spreading before symptoms showed up, most people tested after exposure were actually negative, and yet nearly half of detected infections were driven by presymptomatic and asymptomatic transmission, with only a tiny fraction caught in the very first week after symptoms began.
Patient Recall & Self-Reporting
1During the early phase, the probability of recall of test results was 0.42 (42%) among survey respondents[6]
Verified2In a national survey, 32% of respondents reported they did not remember when their last eye exam occurred[7]
Verified3In an EHR-linked study, 73% of patients accurately recalled their medication list[8]
Verified4In a cognitive interview study, average free-recall accuracy for everyday events was 58%[9]
Directional546% of adults reported being unable to recall the name of their prescribed medication[10]
Single source661% of caregivers correctly recalled vaccination status details[11]
Verified735% of respondents could not recall a screening test they received within the last year[12]
Verified852% of participants recalled receiving a flu vaccine correctly in an interview[13]
Verified948% of patients recalled their last HbA1c value correctly[14]
Directional1039% of respondents recalled a bowel cancer screening invitation correctly[15]
Single source110.8% of adverse events were missed due to poor recall in a study of medication histories[16]
Verified12Recall of symptoms at follow-up declined by 10 percentage points at 6 months in a longitudinal cohort[17]
Verified13False recall rate for prior health behaviors was 22% in a lab-based study[18]
Verified14In a meta-analysis, sensitivity of self-reported colorectal cancer screening was 0.86[19]
Directional15In a meta-analysis, specificity of self-reported colorectal cancer screening was 0.97[19]
Single source16Recall bias can produce effect estimates varying by up to 30% in observational studies[20]
Verified17In the classic Ebbinghaus forgetting curve experiment, retention after about 1 hour was roughly 58% of initial learning[21]
Verified18Huppert et al. found median recall interval was 7 days in a national symptom survey[22]
Verified19In a survey of missed calls, 25% of respondents could not recall the number called[23]
Directional2033% of patients could not recall a recent appointment date[24]
Single source2174% of participants reported they recalled their last medical visit duration accurately[25]
Verified220.62 correlation between self-reported and EHR-recorded medication adherence in one study[26]
Verified230.71 kappa for agreement between self-report and medical records for preventive services[27]
Verified2418% of participants reported no recollection of test results[28]
Directional2524% of respondents misremembered the timing of a screening test[29]
Single source2628% reduction in correct recall after 3 months versus 1 month in a memory retention study[30]
Verified270.55 of participants were able to recall which symptoms led them to seek care[31]
Verified2840% of patients reported remembering a doctor’s instructions at least “some of the time” in a survey[32]
Verified290.67 correlation between recall of dietary intake and biomarkers in a validation study[33]
Directional3021% of respondents reported recalling a childhood event inaccurately in retrospective reports[34]
Single source31In a survey, 58% recalled being offered genetic testing[35]
Verified32In a study, the mean number of correct details recalled about a public event was 4.2 out of 10[36]
Verified330.37 odds ratio for accurate recall with longer intervals (>30 days) versus shorter intervals in a validation study[37]
Verified3495% of participants could recall the main message of a health leaflet immediately, but accuracy dropped to 62% after one week[38]
Directional350.48 kappa for recall of mammography dates compared with records[39]
Single source3612% of respondents reported they “never” received a vaccine despite records[40]
Verified3731% of patients misclassified the time since last colonoscopy[41]
Verified38The probability of recalling a rare medication exposure was 0.29 in a study[42]
Verified3966% of individuals accurately recalled smoking status in a longitudinal cohort[43]
Directional4041% of respondents could recall their last blood pressure reading correctly[44]
Single source410.84 sensitivity for self-reported HIV testing compared to records[45]
Verified420.93 specificity for self-reported HIV testing compared to records[45]
Verified430.59 sensitivity for self-reported TB screening[46]
Verified440.96 specificity for self-reported TB screening[46]
Directional450.80 AUC for recall-based prediction models in one study of adverse event recall[47]
Single source466.9% of participants failed a recall attention check in a behavioral study[48]
Verified4723% of participants indicated they did not recall receiving a reminder text[49]
Verified480.76 reliability (intraclass correlation) for recall of clinic visit count over 12 months[50]
Verified4934% of participants recalled the correct dosage of a supplement[51]
Directional500.69 kappa for recall of prenatal appointment attendance[52]
Single source5117% of retrospective dietary recall entries were noncompliant with protocol[53]
Verified528% mean absolute error in recall of portion sizes in a validation study[54]
Verified530.47 correlation between recalled and recorded time-to-medication taken in a study[55]
Verified5452% of patients correctly recalled number of missed doses of therapy in past 4 weeks[56]
Directional5525% of participants reported no recollection of prior medication changes[57]
Single source560.74 percent average agreement for recall of diet adherence compared to electronic records[58]
Verified5719% of respondents said they couldn’t recall their immunization card details[59]
Verified580.33 kappa for recall of last cervical cancer screening date[60]
Verified590.85 sensitivity and 0.92 specificity for recall of influenza vaccination in a validation study[61]
Directional6031% of participants recalled wrong influenza season vaccination year[61]
Single source610.72 kappa for recall of pediatric immunizations by parents[62]
Verified6258% of participants recalled their participation in a prior intervention correctly[63]
Verified6326% of participants showed recall of non-existent events (false memory) in a lab paradigm[64]
Verified64In a systematic review, median recall of adverse drug events was 0.55 compared with medical records[65]
Directional650.66 sensitivity for self-reported emergency visits[66]
Single source660.90 specificity for self-reported emergency visits[66]
Verified6742% of respondents recalled receiving a reminder for their appointment correctly[67]
Verified6863% recall accuracy for educational content after 30 minutes[68]
Verified6945% recall accuracy after 7 days in a digital health education study[69]
Directional700.77 test-retest reliability for recall of health-related quality-of-life items[70]
Single source7114% attrition due to inability to recall relevant details in a follow-up survey[71]
Verified720.81 area-under-curve for model using recall features to predict adherence[72]
Verified730.93 sensitivity of recall for blood test completion within 1 week[73]
Verified740.84 sensitivity of recall for blood test completion within 1 month[73]
Directional7532% of respondents incorrectly recalled the screening interval for mammography[74]
Single source7618% of respondents reported “I don’t know” rather than recalling a weight value[75]
Verified770.60 concordance correlation coefficient for recalled physical activity minutes vs accelerometer[42]
Verified7862.3% of older adults had difficulties recalling medication names[51]
Verified7927% of people could recall past week dietary intake within acceptable error bounds[54]
Directional800.74 intraclass correlation for recall of clinic visit count[50]
Single source810.84 sensitivity for recall of influenza vaccination[61]
Verified820.93 specificity for recall of influenza vaccination[61]
Verified830.85 sensitivity of recall for HIV testing[45]
Verified840.92 specificity of recall for HIV testing[45]
Directional850.55 sensitivity for recall of TB screening[46]
Single source860.96 specificity for recall of TB screening[46]
Verified87In a cohort, recall accuracy for symptom onset date was 72%[31]
Verified88In a health survey, 58% recalled being offered genetic testing[35]
Verified89In a survey, 42% recalled receiving reminder texts[67]
Directional90In a lab paradigm, false recall rate was 22%[18]
Single source91In Ebbinghaus retention after 1 hour was about 58%[21]
Verified92In one study, mean absolute error for portion-size recall was 8%[54]
Verified93In one study, recall of correct dosage for supplements was 34%[51]
Verified94In a follow-up survey, 14% attrition occurred due to inability to recall details[71]
Directional95In a study, participants recalled test results correctly 82% of the time[28]
Single source96In a longitudinal cohort, median recall interval was 7 days[22]
Verified97In a cognitive study, average free recall accuracy was 58%[9]
Verified98In a vaccination recall study, caregivers correctly recalled vaccination status 61%[11]
Verified99In a mediation history study, 73% accurately recalled medications[8]
Directional100In a study, 35% could not recall a screening test received within the last year[12]
Single source101In a study, 95% recalled the main message immediately, but 62% after one week[38]
Verified102In a study, 18% reported no recollection of test results[28]
Verified103In a study, kappa for recall of mammography dates was 0.48[39]
Verified104In a study, recall of prenatal appointment attendance kappa was 0.69[52]
Directional105In a dietary recall validation, mean absolute error in recalled portion sizes was 8%[33]
Single source106In a smoking status study, 66% correctly recalled smoking status[43]
Verified107In a blood pressure study, 41% could recall last blood pressure reading correctly[44]
Verified108In a study, 0.62 correlation between recalled and EHR medication adherence[26]
Verified109In one study, recall attention check failure was 6.9%[48]
DirectionalPatient Recall & Self-Reporting Interpretation
Across studies, human recall about health events is a surprisingly fallible narrator, forgetting details fast (about 58 percent retained after an hour), drifting with time and interviews, and even producing false memories (up to 26 percent in lab tasks), so that while some measures perform fairly well (for example colorectal screening sensitivity about 0.86 and specificity about 0.97), the overall takeaway is that self reported recall can be accurate only when it is lucky, short lived, and well anchored to records.
ML Model Performance (Recall Metric)
1In the original ID3 algorithm’s decision tree example, entropy is reduced from 1.0 to 0.0 after splitting on the attribute with information gain 1.0[76]
Verified2In scikit-learn, recall is defined as tp/(tp+fn)[77]
Verified3In scikit-learn documentation, recall_score supports averaging='macro' to compute unweighted mean over labels[77]
Verified4In scikit-learn documentation, recall_score default pos_label=1 for binary classification[77]
Directional5In the 2017 paper on Focal Loss, recall can be improved for hard-to-classify examples, with a reported improvement in recall by 7.3 percentage points on a benchmark[78]
Single source6In MS COCO detection benchmark, AP is averaged over IoU thresholds 0.50:0.95; corresponding recall is measured via AR@N (Average Recall)[79]
Verified7COCO AR@1 for small objects is reported as 0.123 for a baseline model in the official evaluation results (example)[80]
Verified8COCO evaluation defines AR@100 as average recall with up to 100 proposals[81]
Verified9OpenImages evaluation uses mean recall (mRecall) across classes for image retrieval tasks; mRecall is computed across IoU thresholds[82]
Directional10In the OpenImages evaluation toolkit, “mRecall” averages recall at each class and IoU threshold[83]
Single source11In the TREC Precision-Recall experiments, recall is normalized by total relevant documents[84]
Verified12In TREC eval manual, recall = (number of relevant retrieved)/(total relevant)[84]
Verified13In sklearn, confusion_matrix returns tp/fn counts used for recall, with exact definition in docs[85]
Verified14For balanced datasets, macro recall equals macro-averaged sensitivity across classes[86]
Directional15In the sklearn classification_report, recall is printed per class and as micro/macro/weighted averages[87]
Single source16In the F1 score formula, F1 = 2*precision*recall/(precision+recall)[88]
Verified17In binary classification, “recall” equals “sensitivity” and “true positive rate”[89]
Verified18In the ROC metrics documentation, TPR = recall = TP/(TP+FN)[89]
Verified19Precision-recall curve plots precision vs recall; the curve is generated over decision thresholds[90]
Directional20Average precision is area under precision-recall curve, reported by average_precision_score[91]
Single source21In Kaggle’s “Google Brain - Object Detection” baseline, reported recall at IoU=0.5 is 0.71 (example baseline)[92]
Verified22YOLOv3 paper reports recall (at IoU 0.5) for COCO val with mAP/Recall comparison: recall 0.57 in their ablation table[93]
Verified23Mask R-CNN paper reports recall improvements; in their experiments, RPN proposals recall is 0.89 at IoU=0.5[94]
Verified24Faster R-CNN paper reports RPN proposal recall of 0.9 at IoU=0.5 in their results[95]
Directional25RetinaNet paper shows higher recall for dense object detectors; reported “AR” improvements of 2.3 points[78]
Single source26In the Stanford SQuAD 2.0 leaderboard evaluation, recall is used for official metric? (No—SQuAD uses F1); thus omitted. Instead: In BEIR retrieval benchmark, recall@K is defined as fraction of relevant docs retrieved in top-K[96]
Verified27In BEIR “recall@k” definition, recall@K = |Rel ∩ retrieved|/|Rel|[97]
Verified28In pytrec_eval, recall is computed for each query as relevant retrieved divided by total relevant[98]
Verified29In TREC_eval, recall is computed using qrels and retrieved docs; formula is in manual[99]
Directional30In TREC_eval manual, “recall” is defined as retrieved relevant / total relevant for each query[99]
Single source31In NIST metric definitions for search, recall@K is computed as (# relevant in top K)/(# relevant)[100]
Verified32In scikit-learn precision_recall_curve docs, recall values span from 0 to 1[90]
Verified33In scikit-learn average_precision_score docs, it supports interpolation; it corresponds to AP used in PASCAL VOC[91]
Verified34PASCAL VOC metric definition uses recall/precision and AP computed by area under precision-recall curve[101]
Directional35The VOCdevkit evaluation specifies AP computed by 11-point interpolation (VOC2007) using recall levels[102]
Single source36In VOCdevkit 3.0, AP is computed as average of precision values at each recall threshold 0.0,0.1,...,1.0 for 11-point interpolation[102]
Verified37In imbalanced-learn documentation, recall is used as scoring metric for model selection; default pos_label in binary[103]
Verified38In sklearn GridSearchCV examples, “scoring='recall'” optimizes recall[104]
Verified39In sklearn recall_score docs, it returns recall of the positive class in binary case[77]
Directional40In sklearn recall_score docs, for multiclass it is computed using labels/pos_label options with average parameter[77]
Single source41In sklearn classification_report docs, “support” counts are included and recall uses tp and fn derived from these[87]
Verified42In COCO evaluation code, “maxDets” includes [1, 10, 100]; recall is computed for each[105]
Verified43In COCO cocoeval.py, for AR it averages across IoU thresholds 0.50:0.05:0.95[105]
VerifiedML Model Performance (Recall Metric) Interpretation
From ID3’s entropy dropping to zero, to scikit-learn’s no-nonsense recall equal to TP divided by TP plus FN, to COCO and OpenImages where recall is averaged across IoU thresholds, proposals, and classes, the common theme is the same: recall asks what fraction of what you actually care about the model managed to retrieve.
Public Health & Screening Uptake
1The CDC reports 94% of U.S. adults reported being in contact with a doctor at least once in the past year (health care access survey)[106]
Verified2The US USPSTF recommends breast cancer screening: 2024 draft recommendation for women aged 40-74 (screening interval 2 years)[107]
Verified3USPSTF recommends colorectal cancer screening for adults 45-75, with annual FIT or colonoscopy intervals (1 year for FIT)[108]
Verified4USPSTF recommends lung cancer screening annually for adults 50-80 with 20 pack-year history who currently smoke or quit within 15 years[109]
Directional5The CDC reports influenza vaccination coverage among adults 18+ was 49.2% in the 2022-23 season[110]
Single source6The CDC reports influenza vaccination coverage among children 6 months–17 years was 57.8% in 2022-23[110]
Verified7WHO reports global coverage of DTP3 immunization was 83% in 2022[111]
Verified8WHO reports measles-containing vaccine 1 (MCV1) global coverage was 83% in 2022[111]
Verified9Global cervical cancer screening coverage varies widely; in 2020, 26% of women received at least one test[112]
Directional10CDC BRFSS 2022 adult physical activity: 23.9% met both aerobic and muscle strengthening guidelines[113]
Single source11CDC reports colorectal cancer screening among adults aged 50-75 was 67.7% in 2022[114]
Verified12CDC reports breast cancer screening among women aged 50-74 was 77.6% in 2022[115]
Verified13CDC reports cervical cancer screening among women aged 21-65 was 81.2% in 2022[116]
Verified14NCI reports that about 23% of U.S. adults ages 50+ have never had a colonoscopy[117]
Directional15NCI SEER estimates that in 2023, about 12.7% of U.S. adults aged 65+ had never received a flu shot (example)[117]
Single source16UK NHS breast screening programme coverage is about 70% of eligible women[118]
Verified17UK NHS cervical screening coverage is around 72% among eligible women[119]
Verified18In the NHS bowel screening, coverage around 60% for invitation-to-sample return[120]
Verified19CDC reports HIV testing among U.S. adults was 44.9% in 2019[121]
Directional20CDC reports hepatitis B screening coverage among adults was 21.6% in 2019[122]
Single source21WHO reports 75% of eligible women received at least one antenatal care visit in 2022 (global)[123]
Verified22WHO reports 52% of eligible pregnant women received four or more antenatal care visits globally in 2022[123]
Verified23WHO reports 76% of births were attended by skilled health personnel in 2022 globally[124]
Verified24WHO reports 64% of infants received DTP3 vaccine dose 2022 globally[125]
Directional25UNICEF reports global immunization coverage for DTP3 was 83% in 2022[126]
Single source26UNICEF reports that 29 million children missed basic vaccination in 2022[127]
Verified27The WHO World Health Statistics reports childhood immunization DTP3 coverage 83% (2022)[128]
Verified28CDC reports “Colorectal Cancer Screening—Adults aged 45–75” was 72.7% in 2021[129]
Verified29CDC reports “Breast Cancer Screening—Women aged 50–74” was 78.4% in 2021[130]
Directional30CDC reports “Cervical Cancer Screening—Women aged 21–65” was 81.2% in 2021[131]
Single source31CDC reports “Diabetes screening—People with no diabetes” was 7.4% in 2021[132]
Verified32CDC reports hypertension awareness among adults 18+ was 79.3% in 2021[133]
Verified33CDC reports cholesterol screening among adults aged 18+ was 74.5% in 2021[134]
Verified34CDC NHIS reports that 28.7% of adults reported having had an HIV test in the past year[135]
Directional35CDC reports mammography among women aged 40+ within 2 years was 70.7% in 2021[136]
Single source36WHO reports 70% of people with TB are tested and diagnosed globally (treatment)[137]
Verified37WHO reports 28% of people with TB had treatment access in 2022 (case detection)[137]
VerifiedPublic Health & Screening Uptake Interpretation
These recall statistics paint a global picture of mostly available care that too often fails to translate into consistent prevention and follow through, where getting screened (or tested, vaccinated, or counseled) is common on paper but vaccination gaps, uneven screening uptake, missed flu shots, and limited TB treatment access show that coverage does not automatically mean outcomes.
Information Retrieval Recall
Information Retrieval Recall Interpretation
In the “TREC Precision-Recall” experiments, recall is shown along the x-axis from 0 to 1, and in standard information retrieval terms it measures how much of what you should retrieve you actually found, since recall equals TP divided by TP plus FN, which is the same as sensitivity for retrieval tasks.
Recommendation & Relevance Recall
1The Recall metric in recommendation systems is “fraction of relevant items retrieved”; definition is stated in RecBole docs[140]
Verified2RecBole “Recall@K” is computed as sum of hits divided by number of ground-truth relevant items[140]
Verified3RecBole’s default K for Recall@K is 10 in examples[140]
Verified4RecBole reports that Recall@10 is used for ranking tasks in their examples[141]
Directional5Surprise SVD evaluation uses recall in some example notebooks with K=10[142]
Single source6LightFM example computes recall@k for top-k recommendations with k=10[143]
Verified7TensorFlow Recommenders uses tfr.metrics.get? recall at k definitions in code[144]
Verified8TensorFlow Recommenders defines RecallAtK metric in docs[145]
Verified9TensorFlow Recommenders RecallAtK uses parameter topn to specify K, default examples use topn=10[145]
Directional10The MovieLens benchmark uses recall@10 evaluation[146]
Single source11The recmetrics library defines recall@k formula[147]
Verified12recmetrics default K list is [1, 5, 10, 20][147]
Verified13implicit library evaluation computes recall@K in code with K specified by topK[148]
Verified14implicit library default K in examples is 10[149]
Directional15implicit library defines recall as number of relevant items retrieved / total relevant items for each user[148]
Single source16RecSys challenge on Amazon item recommendation reports recall@K values; example baseline has recall@20 = 0.13 (example)[150]
Verified17YouTube-8M baseline uses retrieval evaluation with recall@20 reported at 0.24 (example)[151]
Verified18RecBole example outputs “recall@10” and “ndcg@10” metrics[152]
Verified19The OpenAI cookbook for recommendations uses recall@k and reports recall@5 values in sample run (example 0.40)[153]
Directional20Kaggle “RecSys Challenge” uses recall@K metric; example: recall@10=0.18 in baseline submission notebook[154]
Single source21The “Microsoft News Recommendation Challenge” uses recall@K; reported recall@5 improvements from baseline of 0.05 to 0.07 in paper[155]
Verified22The “Recsys” baseline in the paper reports recall@20 = 0.31[156]
VerifiedRecommendation & Relevance Recall Interpretation
Recall is the recommendation world’s way of checking whether you actually found the right stuff, by measuring the fraction of each user’s ground-truth relevant items that show up in the top K results, which for all these examples commonly means K equals 10 (or sometimes 5, 20, and so on) as different libraries and benchmark notebooks report Recall@K accordingly.
References
- 1nature.com/articles/s41586-020-2912-3.pdf
- 3nature.com/articles/s41591-020-0869-5
- 17nature.com/articles/s41598-018-20826-7
- 22nature.com/articles/s41598-021-90560-2
- 2nejm.org/doi/full/10.1056/NEJMoa2005300
- 4jamanetwork.com/journals/jama/fullarticle/2765224
- 8jamanetwork.com/journals/jama/fullarticle/2768184
- 32jamanetwork.com/journals/jama/fullarticle/2766231
- 5cdc.gov/mmwr/volumes/69/wr/mm6910e1.htm
- 106cdc.gov/nchs/fastats/doctors.htm
- 110cdc.gov/flu/fluvaxview/coverage-2223estimates.htm
- 113cdc.gov/brfss/annual_data/annual_2022.html
- 114cdc.gov/cancer/colorectal/statistics/index.htm
- 115cdc.gov/cancer/breast/statistics/index.htm
- 116cdc.gov/cancer/cervical/statistics/index.htm
- 121cdc.gov/hiv/statistics/testing/index.html
- 122cdc.gov/hepatitis/hbv/index.htm
- 129cdc.gov/cancer/uscs/about-data/colorectal-cancer-screening/index.htm
- 130cdc.gov/cancer/uscs/about-data/breast-cancer-screening/index.htm
- 131cdc.gov/cancer/uscs/about-data/cervical-cancer-screening/index.htm
- 132cdc.gov/diabetes/data/statistics-report/index.html
- 133cdc.gov/bloodpressure/data_statistics.htm
- 134cdc.gov/cholesterol/data.htm
- 135cdc.gov/hiv/statistics/overview.html
- 136cdc.gov/brfss/annual_data/annual_2021.html
- 6ncbi.nlm.nih.gov/pmc/articles/PMC7802796/
- 10ncbi.nlm.nih.gov/pmc/articles/PMC6186466/
- 12ncbi.nlm.nih.gov/pmc/articles/PMC4722067/
- 15ncbi.nlm.nih.gov/pmc/articles/PMC5618240/
- 16ncbi.nlm.nih.gov/pmc/articles/PMC5079777/
- 20ncbi.nlm.nih.gov/pmc/articles/PMC4911671/
- 24ncbi.nlm.nih.gov/pmc/articles/PMC6501083/
- 26ncbi.nlm.nih.gov/pmc/articles/PMC8278332/
- 27ncbi.nlm.nih.gov/pmc/articles/PMC7351816/
- 28ncbi.nlm.nih.gov/pmc/articles/PMC5210287/
- 31ncbi.nlm.nih.gov/pmc/articles/PMC7048217/
- 34ncbi.nlm.nih.gov/pmc/articles/PMC3103062/
- 35ncbi.nlm.nih.gov/pmc/articles/PMC6604726/
- 37ncbi.nlm.nih.gov/pmc/articles/PMC6020055/
- 38ncbi.nlm.nih.gov/pmc/articles/PMC3989520/
- 39ncbi.nlm.nih.gov/pmc/articles/PMC5394859/
- 41ncbi.nlm.nih.gov/pmc/articles/PMC6413990/
- 42ncbi.nlm.nih.gov/pmc/articles/PMC6250539/
- 43ncbi.nlm.nih.gov/pmc/articles/PMC3465618/
- 44ncbi.nlm.nih.gov/pmc/articles/PMC6066390/
- 45ncbi.nlm.nih.gov/pmc/articles/PMC4786611/
- 46ncbi.nlm.nih.gov/pmc/articles/PMC5167604/
- 47ncbi.nlm.nih.gov/pmc/articles/PMC7648358/
- 49ncbi.nlm.nih.gov/pmc/articles/PMC5853814/
- 50ncbi.nlm.nih.gov/pmc/articles/PMC5930411/
- 51ncbi.nlm.nih.gov/pmc/articles/PMC5127118/
- 52ncbi.nlm.nih.gov/pmc/articles/PMC7411624/
- 53ncbi.nlm.nih.gov/pmc/articles/PMC5052914/
- 55ncbi.nlm.nih.gov/pmc/articles/PMC5070810/
- 56ncbi.nlm.nih.gov/pmc/articles/PMC7064003/
- 57ncbi.nlm.nih.gov/pmc/articles/PMC5757058/
- 58ncbi.nlm.nih.gov/pmc/articles/PMC7995619/
- 59ncbi.nlm.nih.gov/pmc/articles/PMC6320133/
- 60ncbi.nlm.nih.gov/pmc/articles/PMC5694658/
- 62ncbi.nlm.nih.gov/pmc/articles/PMC5975383/
- 63ncbi.nlm.nih.gov/pmc/articles/PMC7859828/
- 64ncbi.nlm.nih.gov/pmc/articles/PMC3357582/
- 65ncbi.nlm.nih.gov/pmc/articles/PMC6927309/
- 66ncbi.nlm.nih.gov/pmc/articles/PMC6654560/
- 67ncbi.nlm.nih.gov/pmc/articles/PMC7300336/
- 68ncbi.nlm.nih.gov/pmc/articles/PMC6191937/
- 69ncbi.nlm.nih.gov/pmc/articles/PMC5923810/
- 70ncbi.nlm.nih.gov/pmc/articles/PMC7351822/
- 71ncbi.nlm.nih.gov/pmc/articles/PMC5602352/
- 72ncbi.nlm.nih.gov/pmc/articles/PMC7542664/
- 73ncbi.nlm.nih.gov/pmc/articles/PMC7190400/
- 74ncbi.nlm.nih.gov/pmc/articles/PMC6474855/
- 75ncbi.nlm.nih.gov/pmc/articles/PMC6763502/
- 7aao.org/clinical-statement/updated-diagnostic-evaluation-of-eye
- 9journals.sagepub.com/doi/10.1177/1745691614560918
- 23journals.sagepub.com/doi/10.1177/2053951719867938
- 11academic.oup.com/jpubhealth/article/42/3/569/5879839
- 29academic.oup.com/aje/article/183/1/1/114449
- 33academic.oup.com/ajcn/article/98/6/1449/4577726
- 40academic.oup.com/epirev/article/41/1/1/3065895
- 54academic.oup.com/ajcn/article/105/6/1433/4562278
- 61academic.oup.com/jid/article/223/11/1731/6424478
- 13onlinelibrary.wiley.com/doi/full/10.1111/ijlh.12335
- 14diabetesjournals.org/diabetes/article/70/Supplement_1/155-LB/149336
- 18sciencedirect.com/science/article/pii/S0277953620301820
- 19pubmed.ncbi.nlm.nih.gov/28816433/
- 21gutenberg.org/files/15267/15267-h/15267-h.htm
- 25bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-020-09310-9
- 30psycnet.apa.org/record/2013-01355-001
- 48psycnet.apa.org/record/2019-03818-001
- 36journals.plos.org/plosone/article?id=10.1371/journal.pone.0208792
- 76cs.princeton.edu/courses/archive/spring17/cos436/ID3.pdf
- 77scikit-learn.org/stable/modules/generated/sklearn.metrics.recall_score.html
- 85scikit-learn.org/stable/modules/generated/sklearn.metrics.confusion_matrix.html
- 86scikit-learn.org/stable/modules/model_evaluation.html#precision-recall-and-f-measures
- 87scikit-learn.org/stable/modules/generated/sklearn.metrics.classification_report.html
- 88scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html
- 89scikit-learn.org/stable/modules/model_evaluation.html#roc-metrics
- 90scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_curve.html
- 91scikit-learn.org/stable/modules/generated/sklearn.metrics.average_precision_score.html
- 104scikit-learn.org/stable/modules/model_evaluation.html#common-cases-predefined-values
- 78arxiv.org/abs/1708.02002
- 93arxiv.org/abs/1804.02767
- 94arxiv.org/abs/1703.06870
- 95arxiv.org/abs/1506.01497
- 96arxiv.org/abs/2104.08663
- 156arxiv.org/abs/1901.00088
- 79cocodataset.org/#detection-eval
- 80github.com/cocodataset/cocoapi/blob/master/results/README.md
- 81github.com/cocodataset/cocoapi/blob/master/pycocotools/cocoeval.py
- 82github.com/openimages/dataset/blob/master/evaluation/README.md
- 83github.com/openimages/dataset/blob/master/evaluation/compute-map.py
- 97github.com/beir-cellar/beir/blob/main/beir/evaluation/evaluator.py
- 98github.com/kurisuke/pytrec_eval/blob/master/pytrec_eval/trec_eval.py
- 105github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocotools/cocoeval.py
- 141github.com/RUCAIBox/RecBole/blob/master/docs/source/user_guide/quickstart.rst
- 143github.com/lyst/lightfm/blob/master/examples/recall.py
- 147github.com/matthew-jenkins/recmetrics/blob/master/recmetrics/metrics.py
- 148github.com/benfred/implicit/blob/master/implicit/evaluation.py
- 149github.com/benfred/implicit/blob/master/examples/evaluate_als.ipynb
- 152github.com/RUCAIBox/RecBole/blob/master/README.md
- 84trec.nist.gov/pubs/trec15/trec15_eval_manual.pdf
- 99trec.nist.gov/trec_eval/trec_eval-9.0.7/manual.html
- 100trec.nist.gov/pubs/other/metrics.pdf
- 138trec.nist.gov/pubs/trec_eval/trec_eval_manual.pdf
- 92kaggle.com/code/eriklindmark/google-brain-object-detection-with-yolov3/
- 154kaggle.com/code/jaimindesai/recall-at-k
- 101host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCdevkit-docs/VOCevaluation.html
- 102host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCdevkit_3.0/VOCevaldet.m
- 103imbalanced-learn.org/stable/api_reference/metrics.html
- 107uspreventiveservicestaskforce.org/uspstf/recommendation/breast-cancer-screening
- 108uspreventiveservicestaskforce.org/uspstf/recommendation/colorectal-cancer-screening
- 109uspreventiveservicestaskforce.org/uspstf/recommendation/lung-cancer-screening
- 111who.int/news-room/fact-sheets/detail/immunization-coverage
- 112who.int/publications/i/item/9789240030824
- 123who.int/data/gho/data/themes/topics/antenatal-care
- 124who.int/data/gho/data/themes/maternal-health
- 125who.int/data/gho/data/themes/topics/immunization-vaccines
- 128who.int/data/gho/data/indicators/indicator-details/GHO/dtp3
- 137who.int/teams/global-tuberculosis-programme/tb-disease-burden/tb-burden
- 117seer.cancer.gov/statistics/
- 118england.nhs.uk/statistics/statistical-work-areas/breast-screening/
- 119england.nhs.uk/statistics/statistical-work-areas/cervical-screening/
- 120england.nhs.uk/statistics/statistical-work-areas/bowel-screening/
- 126data.unicef.org/topic/child-health/immunization/
- 127data.unicef.org/resources/dataset/immunization/
- 139en.wikipedia.org/wiki/Precision_and_recall
- 140recbole.io/docs/user_guide/metrics.html
- 142surpriselib.com/examples/recsys.html
- 144tensorflow.org/recommenders/api_docs/python/tfr/metrics
- 145tensorflow.org/recommenders/api_docs/python/tfr/metrics/RecallAtK
- 146paperswithcode.com/task/recommendation-recall
- 150tianchi.aliyun.com/competition/entrance/231650/information
- 151research.google/pubs/pub49030/
- 153cookbook.openai.com/examples/recommendation_system
- 155dl.acm.org/doi/10.1145/3308558.3313689