Process / pipelineDeneysel desen

آزمون A/B تطبیقی — آزمون A/B تطبیقی

آزمون A/B تطبیقی یک طراحی آزمایشی است که به صورت پویا ترافیک یا شرکت‌کنندگان را در طول خود آزمایش به سمت واریانت‌های با عملکرد بهتر تخصیص مجدد می‌دهد، به جای اینکه تخصیص‌ها را تا پایان ثابت نگه دارد. این آزمون با بهره‌گیری از الگوریتم‌های چندبازو مانند نمونه‌برداری تامپسون (Thompson Sampling) یا کران بالای اطمینان (UCB)، بین کاوش واریانت‌های نامطمئن و بهره‌برداری از واریانت‌هایی که عملکرد برتری از خود نشان داده‌اند، تعادل برقرار می‌کند و معمولاً نتایج تجمعی بالاتری را به همراه دارد، در حالی که همچنان نتایج استنباطی معتبری تولید می‌کند.

یافتن موضوع با PaperMindبه‌زودیویدیوبه‌زودیDownload slides

مطالعهٔ کامل روش

ویژهٔ اعضا

برای خواندن این بخش با حساب رایگان وارد شوید.

ورود

Method map

The neighbourhood of related methods — select a node to explore.

منابع

  1. Russo, D., Van Roy, B., Kazerouni, A., Osband, I., & Wen, Z. (2018). A Tutorial on Thompson Sampling. Foundations and Trends in Machine Learning, 11(1), 1–96. DOI: 10.1561/2200000070
  2. Offer-Westort, M., Coppock, A., & Green, D. P. (2021). Adaptive Experimental Design: Prospects and Applications in Political Science. American Journal of Political Science, 65(4), 826–844. DOI: 10.1111/ajps.12597

نحوهٔ استناد به این صفحه

ScholarGate. (2026, June 3). Adaptive A/B Testing. ScholarGate. https://scholargate.app/fa/experimental-design/adaptive-ab-test

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

Compare side by side

ارجاع‌شده در

ScholarGateAdaptive A/B test (Adaptive A/B Testing). بازیابی‌شده در 2026-06-15 از https://scholargate.app/fa/experimental-design/adaptive-ab-test · مجموعه‌داده: https://doi.org/10.5281/zenodo.20539026