VolgendeVolgende gids
Klasse-onbalans en herbemonstering
Technisch
Technische GIDS
Bootstrap-resampling schat de onzekerheid in door herhaaldelijk nieuwe steekproeven te trekken, met vervanging, uit de gegevens die u hebt waargenomen en een statistiek opnieuw te berekenen.
Luister naar deze gidsAI-verteld · ~3 min
It can help describe how a mean or model-evaluation metric varies, provided the resampling design matches how the observations are related.
A statistic calculated from one dataset is only an estimate. Another sample from the same population could produce a different mean, correlation or evaluation score. The bootstrap approximates aspects of that sampling variation using the observations already available. In the ordinary nonparametric bootstrap for independent observations, draw a sample of the original size with replacement. Each draw can select any original observation again, so some observations appear more than once and others are omitted. Calculate the statistic on that sample. Repeat the process many times to produce a distribution of bootstrap statistics. For an arithmetic illustration, the values 2, 4 and 9 have a mean of five. A possible resample is 2, 2 and 9, whose mean is about 4.33. One resample does not establish an uncertainty interval. The variation across many appropriately constructed resamples is what makes the procedure useful. A percentile interval takes endpoints from percentiles of the bootstrap distribution. Other methods, including the bias-corrected and accelerated method supported by SciPy, adjust the construction differently. The chosen method and its assumptions matter, particularly for small samples or difficult statistics. For paired data, preserve the pairing. When comparing model predictions on the same cases, resample each case with its outcome and both predictions. Breaking those relationships changes the question being evaluated. The bootstrap cannot create evidence about people or conditions missing from the original sample. It also does not automatically correct selection bias or dependence between observations. Repeated measurements may require resampling groups; time series may require a suitable block method. Choose the resampling unit to reflect the data-generating process before interpreting the apparent precision.
Starting from the illustrative values 2, 4 and 9, one bootstrap sample might be 2, 2 and 9. It has the same size as the original sample, includes a repeated observation and has a mean of 13 divided by three.
To compare two models on the same test cases, an analyst resamples case identifiers and carries both models' predictions and the true outcome together. Each resample yields a difference in their scores.
A study has several records per participant. The analyst considers resampling whole participants rather than treating every record as independent.
A team uses scipy.stats.bootstrap to estimate uncertainty in a statistic. It records the interval method and random seed and checks whether the resulting interval is stable enough for its intended use.
Evaluation reports could become more useful by recording the resampling unit, interval method and scope of uncertainty alongside the resulting bounds. A reader should be able to tell whether an interval describes a fixed model on new cases or a process that includes refitting. Teams can also preserve resampled indices and software versions when reproducibility matters. The next improvement is often a better evaluation sample rather than more bootstrap repetitions. Understanding which sources of uncertainty are included makes the interval more actionable than presenting narrow bounds without their assumptions.
Gratis nieuwsbrief
Drie verifieerde AI-verhalen elke weekdagmorgen, geschreven in eenvoudig Engels. Gratis voor altijd, geen advertenties.
Eén e-mail elke weekdag. Afmelden met één klik. We verkopen of delen je adres nooit.
Test jezelf
Directe feedback op elke antwoord, en een deelbare certificaat met een verifieerbare ID zodra je een cursus hebt afgerond.
Ondersteun gratis AI-onderwijs. AI Understanding is een 501(c)(3) non-profit — geen advertenties, geen betaalmuur, nooit. Doe een donatie
Bootstrap resampling estimates uncertainty by repeatedly drawing new samples, with replacement, from the data you observed and recalculating a statistic. It can help describe how a mean or model-evaluation metric varies, provided the resampling design matches how the observations are related.
A bootstrap sample has the original size and is drawn with replacement, allowing repeated observations.
The three resampled values sum to 13, so their mean is 13 divided by three, approximately 4.33.
Preserving the case-level pairing allows the resampled score difference to compare predictions for the same observations.
The percentile method uses quantiles of the resampled statistics, with 2.5% left in each tail for this interval.
More repetitions make the simulation more stable but do not provide new independent observations or repair sampling problems.
Blijf leren
Er zijn meer handleidingen voor dit onderwerp geselecteerd
VolgendeVolgende gids
Klasse-onbalans en herbemonstering
Technisch