F1
The harmonic mean of precision and recall, measuring classification quality at the event-count level. It assumes all errors carry equal cost and is suited to balanced classification scenarios.
Core Facts
Timeline (last 90 days)
IPIP persona combinations improved the semantic parser's F1 in all four matched comparisons (Holm-corrected p<.001), and all three persona combinations had positive main effects across the full 2³ grid.
There is a mathematical relationship between the Jaccard Index and F1: Jaccard = F1 / (2 - F1), and the two can be converted into each other.
IDP scoring should not report only a single metric; at minimum it should give Precision, Recall, and F1 together and state the counting rules for blanks, NULLs, checkboxes, and unreadable values.
The 1978 Brabham BT46B F1 car used a fan-car design to win one race before being banned.
Precision = true positives / (true positives + false positives), recall = true positives / (true positives + false negatives); F1 is the harmonic mean of the two, implicitly assuming they are equally important.
When scanning thresholds in steps of 0.01, precision rises while recall falls, the two cross near a threshold of 0.63 (P≈0.733, R≈0.725), and the F1 peak is at a threshold of 0.41 (F1≈0.758).
For a binary classification feature, one should choose a threshold on a labeled holdout set reflecting the production distribution's skew, based on acceptable false-positive and false-negative costs, report F1 at that threshold, and keep AUC on the scored samples for comparing models before locking in the cutoff.
For multiclass routing, one should use per-class F1 and choose an averaging method that can be justified by error costs.
Amount-weighted recall weights fraud samples by the amounts involved and measures the proportion of fraud amount intercepted; it may select a different model than F1.
An evaluation culture centered on single metrics such as accuracy and F1 may be measuring in the wrong direction when faced with inherently ambiguous tasks.
12 more timeline events
All Facts (20)
An AUC=1 on the ROC curve indicates a perfect classifier, while AUC=0.5 is equivalent to random guessing.
70%Unverified现代F1赛车每辆搭载超过300个传感器，每秒可产生约1500个数据点，单场比赛原始数据量可达数十TB
75%Unverified每场F1比赛会产生数十亿个数据点，涵盖赛车遥测、轮胎磨损、进站策略、天气变化等多个维度
75%UnverifiedNetflix《Drive to Survive》系列自2019年播出后，F1在北美和亚太地区的新增粉丝中，女性观众比例从14%上升至40%
70%UnverifiedIPIP persona combinations improved the semantic parser's F1 in all four matched comparisons (Holm-corrected p<.001), and all three persona combinations had positive main effects across the full 2³ grid.
50%UnverifiedIDP scoring should not report only a single metric; at minimum it should give Precision, Recall, and F1 together and state the counting rules for blanks, NULLs, checkboxes, and unreadable values.
50%UnverifiedThere is a mathematical relationship between the Jaccard Index and F1: Jaccard = F1 / (2 - F1), and the two can be converted into each other.
50%UnverifiedThe 1978 Brabham BT46B F1 car used a fan-car design to win one race before being banned.
50%UnverifiedPrecision = true positives / (true positives + false positives), recall = true positives / (true positives + false negatives); F1 is the harmonic mean of the two, implicitly assuming they are equally important.
50%UnverifiedWhen scanning thresholds in steps of 0.01, precision rises while recall falls, the two cross near a threshold of 0.63 (P≈0.733, R≈0.725), and the F1 peak is at a threshold of 0.41 (F1≈0.758).
50%UnverifiedFor multiclass routing, one should use per-class F1 and choose an averaging method that can be justified by error costs.
50%UnverifiedFor a binary classification feature, one should choose a threshold on a labeled holdout set reflecting the production distribution's skew, based on acceptable false-positive and false-negative costs, report F1 at that threshold, and keep AUC on the scored samples for comparing models before locking in the cutoff.
50%UnverifiedAmount-weighted recall weights fraud samples by the amounts involved and measures the proportion of fraud amount intercepted; it may select a different model than F1.
50%UnverifiedAn evaluation culture centered on single metrics such as accuracy and F1 may be measuring in the wrong direction when faced with inherently ambiguous tasks.
50%Unverified运行在标准ECU硬件上的软件逻辑仍由各车队自行开发和优化
50%UnverifiedF1遥测数据链路让车队工程师能够实时监控超过200个传感器的数据
50%UnverifiedF1赛车软件并不完全按照DO-178C或ISO 26262等民用安全标准开发，其研发周期极短且迭代频繁
50%Unverified现代F1赛车的方向盘集成了数十个按钮与旋钮，背后是ECU、ERS以及实时遥测数据链路
50%UnverifiedThe time for an F1 pit stop tire change has dropped from 67 seconds in the 1950s to 1.8 seconds today.
50%UnverifiedAn F1 car has 16,000 parts, 90% of which are replaced each year, with only 10% carried over to the next season.
50%