Pearl 的 do 算子如何区分「观察」与「干预」？
How does Pearl's do-operator distinguish observation from intervention?
$P(Y\mid X=x)$ 表示在观察到 $X=x$ 时 $Y$ 的分布，而 $P(Y\mid do(X=x))$ 表示主动把 $X$ 设为 $x$ 后的分布。do 算子切断了 $X$ 与其原有原因的联系，从而刻画干预效应，是因果推断区别于纯统计关联的关键。
$P(Y\mid X=x)$ is the distribution of $Y$ when $X=x$ is observed, while $P(Y\mid do(X=x))$ is the distribution after forcibly setting $X=x$. The do-operator severs $X$ from its causes, capturing intervention effects—the crux distinguishing causal inference from mere association.
难度 专业 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Pearl、Glymour 与 Jewell《Causal Inference in Statistics: A Primer》（Wiley）
Pearl, Glymour and Jewell, Causal Inference in Statistics: A Primer, Wiley
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-1-119-18687-8 (2016, Wiley)
#do算子 #干预 #因果
因果推断中的有向无环图（DAG）如何帮助识别可估计的因果效应？
How do directed acyclic graphs (DAGs) help identify estimable causal effects in causal inference?
DAG 用节点表示变量、有向边表示因果关系。通过后门准则可判断是否存在需调整的混杂变量集，使得调整后能无偏估计因果效应；若所有后门路径被阻断，则观测关联可解释为因果。这是 Pearl 因果框架的核心工具。
A DAG represents variables as nodes and causal links as directed edges. The back-door criterion identifies sets of confounders to adjust for so a causal effect is estimated without bias; once all back-door paths are blocked, observed association can be interpreted causally. This is central to Pearl's causal framework.
难度 专业 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Pearl、Glymour 与 Jewell《Causal Inference in Statistics: A Primer》（Wiley）
Pearl, Glymour and Jewell, Causal Inference in Statistics: A Primer, Wiley
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-1-119-18687-8 (2016, Wiley)
#因果图 #后门准则 #混杂
假设检验中的第一类错误与第二类错误分别指什么？
What do Type I and Type II errors refer to in hypothesis testing?
第一类错误（弃真）是原假设实际为真却被拒绝，其概率记为 $\alpha$；第二类错误（取伪）是原假设为假却未被拒绝，概率记为 $\beta$。通常需在两者间权衡，提高检验功效（降低 $\beta$）往往要以增大 $\alpha$ 为代价。
A Type I error (false positive) rejects a true null hypothesis with probability $\alpha$; a Type II error (false negative) fails to reject a false null with probability $\beta$. There is a trade-off: raising power (lowering $\beta$) often requires increasing $\alpha$.
难度 进阶 · 适合 初中(12-15)
CC BY 4.0 · 已核验 · Verified 2026-11-02
来源 · Source
美国数学会(AMS)《数理统计》
American Mathematical Society (AMS), Mathematical Statistics
独立核验依据 · Independent verification: 核验依据 1 ISBN 9787510005343（豆瓣书目条目: 数理统计）
#第一类错误 #第二类错误 #检验功效
贝叶斯推断如何把先验分布与观测数据结合得到后验？
How does Bayesian inference combine a prior distribution with observed data to obtain the posterior?
由贝叶斯公式 $P(\theta\mid D)=\frac{P(D\mid\theta)P(\theta)}{P(D)}$，先验 $P(\theta)$ 与似然 $P(D\mid\theta)$ 结合得到后验 $P(\theta\mid D)$。后验随数据更新，可作为进一步推断与决策的基础；分母 $P(D)$ 仅作归一化常数。
By Bayes' rule $P(\theta\mid D)=\frac{P(D\mid\theta)P(\theta)}{P(D)}$, the prior $P(\theta)$ and likelihood $P(D\mid\theta)$ combine into the posterior $P(\theta\mid D)$. The posterior updates with data and underlies further inference and decisions; $P(D)$ is merely a normalizing constant.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-11-02
来源 · Source
美国数学会(AMS)《数理统计》
American Mathematical Society (AMS), Mathematical Statistics
独立核验依据 · Independent verification: 核验依据 1 ISBN 9787510005343（豆瓣书目条目: 数理统计）
#贝叶斯推断 #后验 #似然
置信区间的正确含义是什么，常见误解有哪些？
What is the correct meaning of a confidence interval, and what are common misinterpretations?
95% 置信区间的含义是：在重复抽样下，这样构造的区间约有 95% 会覆盖真实参数；它并非「参数有 95% 概率落在此区间内」。参数本身是固定常数，区间才是随机的，这是最常见的误解。
A 95% confidence interval means that under repeated sampling about 95% of such intervals would cover the true parameter; it is not that the parameter has a 95% probability of lying in this interval. The parameter is fixed and the interval is random—a common misinterpretation.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-11-02
来源 · Source
美国数学会(AMS)《数理统计》
American Mathematical Society (AMS), Mathematical Statistics
独立核验依据 · Independent verification: 核验依据 1 ISBN 9787510005343（豆瓣书目条目: 数理统计）
#置信区间 #重复抽样 #误解
最大似然估计的基本思想是什么？
What is the basic idea of maximum likelihood estimation?
在给定模型下，选择使观测数据出现概率（似然）最大的参数值作为参数估计，即求解 $\hat\theta=\arg\max_\theta L(\theta;D)$。它直观、渐近有效，且在多数正则条件下具有相合性与渐近正态性。
Under a given model, choose the parameter value that maximizes the probability (likelihood) of the observed data, i.e. $\hat\theta=\arg\max_\theta L(\theta;D)$. It is intuitive, asymptotically efficient, and under regularity conditions is consistent and asymptotically normal.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-11-02
来源 · Source
美国数学会(AMS)《数理统计》
American Mathematical Society (AMS), Mathematical Statistics
独立核验依据 · Independent verification: 核验依据 1 ISBN 9787510005343（豆瓣书目条目: 数理统计）
#最大似然 #似然函数 #相合性