📚 Mastering A-Level Statistics S2: A Computer Science Perspective | 掌握A-Level统计S2：计算机科学视角
In A-Level Mathematics, Unit S2 (Statistics) builds on your earlier understanding of probability and distributions, adding deeper tools such as continuous random variables, Poisson processes and hypothesis testing. For computer science students, these statistical concepts are not just abstract theory—they power reliability modelling, data analysis, simulation and even machine learning algorithms.
在A-Level数学课程中，S2（统计）单元建立在您对概率和分布的初步理解之上，加入了连续随机变量、泊松过程和假设检验等更深入的工具。对于计算机科学学生而言，这些统计概念不仅仅是抽象理论，它们支撑着可靠性建模、数据分析、模拟乃至机器学习算法。
1. Discrete Random Variables and Expectation | 离散随机变量与期望
The expectation E(X) measures the long-run average and is defined as E(X) = Σ x P(X = x).
离散随机变量X取可数个值，每个值的概率介于0和1之间，且所有概率之和等于1。期望E(X)衡量长期平均值，定义为E(X) = Σ x P(X = x)。
E(X) = Σ x P(X = x), Var(X) = E(X²) − [E(X)]²
The variance gives the spread of the distribution. A small variance means that outcomes are consistently close to the mean, which is useful when comparing algorithm running times or packet delays.
方差刻画分布的离散程度。方差小意味着结果稳定接近平均值，这对于比较算法运行时间或数据包延迟非常有用。
In computer systems, the expected number of requests per second is a discrete random variable, and its variance tells us how unpredictable the system load is.
在计算机系统中，每秒请求数是一个离散随机变量，其方差则告诉我们在系统负载方面有多少不可预测性。
2. The Poisson Distribution | 泊松分布
The Poisson distribution is used to model the number of rare events occurring in a fixed interval of time or space. It requires events to happen independently, at a constant average rate λ, and one at a time.
泊松分布用于建模在固定时间或空间间隔内稀有事件发生的次数。它要求事件独立发生、平均速率λ恒定，且每次只发生一个事件。
P(X = k) = exp(−λ) λᵏ / k!, for k = 0, 1, 2, …
For a Poisson variable, the mean and variance are both equal to λ. This property is often tested in S2 exam questions.
对于泊松变量，其均值与方差均等于λ。这一性质在S2考试中经常被考查。
A classic computer science example is the number of corrupted bits in a data packet, or the number of network connection requests arriving at a server in one second.
计算机科学的经典例子包括一个数据包中损坏的比特数，或者一秒钟内到达服务器的网络连接请求数。
3. Poisson Processes in Computer Systems | 计算机系统中的泊松过程
A Poisson process extends the simple Poisson distribution to continuous time. If events occur at rate λ per unit time, the number of events N(t) in an interval of length t has a Poisson distribution with mean λt.
泊松过程将简单泊松分布推广到连续时间。如果事件以每单位时间λ的速率发生，那么在长度为t的区间内的事件数N(t)服从均值为λt的泊松分布。
This model is widely used for network traffic, job arrivals to a CPU queue, or user click streams. It helps engineers simulate and predict system load.
该模型广泛用于网络流量、CPU队列中的作业到达以及用户点击流。它帮助工程师模拟和预测系统负载。
Because the Poisson process is memoryless, the probability of a future event does not depend on how long you have already waited. This property is also shared by the exponential random variable described later.
由于泊松过程具有无记忆性，未来事件发生的概率不依赖于您已经等待了多长时间。这一性质也为后文讨论的指数随机变量所共用。
4. Continuous Random Variables | 连续随机变量
A continuous random variable can take any value in an interval, and its behaviour is described by a probability density function f(x). The total area under f(x) is 1, and the probability of falling between a and b is the integral of f(x) from a to b.
连续随机变量可以在区间内取任意值，其行为由概率密度函数f(x)描述。f(x)下总面积为1，落在a和b之间的概率等于从a到b对f(x)积分。
P(a < X < b) = ∫ₐᵇ f(x) dx, E(X) = ∫ x f(x) dx, Var(X) = ∫ (x − μ)² f(x) dx
Unlike discrete probability, the probability of a single exact value is zero. In computing, this is like measuring response time to the nearest microsecond—the probability of an exact value is effectively zero.
与离散概率不同，单个精确值的概率为零。在计算中，这就像以微秒为单位测量响应时间——精确到某个值的概率实际上为零。
Continuous distributions are used to model file sizes, memory usage, and execution times in performance analysis.
连续分布用于对文件大小、内存使用量和执行时间进行性能分析。
5. The Normal Distribution | 正态分布
The normal distribution is the most important continuous distribution in statistics. It is symmetric and bell-shaped, with parameters μ for the mean and σ² for the variance.
正态分布是统计学中最重要的连续分布。它对称且呈钟形，参数μ表示均值，σ²表示方差。
Z = (X − μ) / σ
To use tables, we standardise X into Z, which has mean 0 and variance 1. In S2 you are expected to find probabilities such as P(Z < 1.5) from standard normal tables.
为了查表，我们将X标准化为Z，其均值为0，方差为1。在S2中，您需要会利用标准正态分布表求如P(Z < 1.5)这样的概率。
In computer science, normal distributions appear in benchmark scores, system latency distributions under many independent factors, and noise models for randomised algorithms.
在计算机科学中，正态分布出现在基准测试分数、多个独立因素作用下的系统延迟分布，以及随机算法中的噪声模型。
6. The Exponential Distribution | 指数分布
The exponential distribution is closely linked to the Poisson process. If events occur according to a Poisson process with rate λ, then the waiting time until the next event has an exponential distribution with parameter λ.
指数分布与泊松过程密切相关。如果事件按速率为λ的泊松过程发生，那么直到下一个事件的等待时间服从参数为λ的指数分布。
f(x) = λ exp(−λx), for x ≥ 0; E(X) = 1/λ, Var(X) = 1/λ²
The mean waiting time is 1/λ, and the distribution has a long tail, meaning very long waits are rare but possible. This is typical of service times in a bus system or the time between two network packets.
平均等待时间为1/λ，分布具有长尾，意味着非常长的等待虽然罕见但可能出现。这类似于公交系统中两次服务之间的时间，或者两个网络数据包之间的时间。
Because of the memoryless property, the conditional probability that X > s + t given X > s is the same as P(X > t). This is useful in cache miss analysis and timeout design.
由于无记忆性，在X > s的条件下X > s + t的条件概率等于P(X > t)。这一性质在缓存未命中分析和超时设计中很有用。
7. Sampling and Sample Distributions | 抽样与样本分布
In statistics, we often take a random sample X₁, X₂, …, Xₙ from a population to estimate parameters. The sample mean X̄ is itself a random variable, and its distribution is called the sampling distribution.
在统计学中，我们常常从总体中抽取随机样本X₁, X₂, …, Xₙ来估计参数。样本均值X̄本身是一个随机变量，其分布称为抽样分布。
E(X̄) = μ, Var(X̄) = σ² / n
If the population is normal, the sample mean is exactly normal. For non-normal populations, the sample mean becomes approximately normal when n is large, thanks to the central limit theorem.
如果总体是正态的，样本均值精确服从正态分布。对于非正态总体，当n很大时，由于中心极限定理，样本均值近似服从正态分布。
In a computer science context, sampling can be used to estimate the average response time of a server without collecting data from every user request.
在计算机科学背景下，抽样可用于估计服务器的平均响应时间，而无需收集每个用户请求的数据。
8. The Central Limit Theorem | 中心极限定理
The central limit theorem states that, for a sufficiently large sample size n, the distribution of the sample mean X̄ is approximately normal with mean μ and variance σ²/n, regardless of the population's original distribution.
中心极限定理指出，当样本量n足够大时，样本均值X̄的分布近似为正态分布，其均值为μ，方差为σ²/n，无论总体原始分布是什么。
X̄ ≈ N(μ, σ²/n) for large n
This lets us make probability statements about sample averages even when we do not know the exact population distribution.
在实践中，n ≥ 30通常就足够了。这使我们即使不知道总体的确切分布，也能对样本平均值进行概率推断。
Examples include averaging many independent measurements of memory latency, or estimating the average size of files in a storage system.
例如对多次独立测量内存延迟取平均，或者估计存储系统中文件的平均大小。
9. Hypothesis Testing | 假设检验
Hypothesis testing is a formal procedure for deciding whether a claim about a population parameter is supported by the sample evidence. We start with a null hypothesis H₀ and an alternative hypothesis H₁.
假设检验是一种正式的程序，用于决定关于总体参数的断言是否得到样本证据的支持。我们从零假设H₀和备择假设H₁开始。
We choose a significance level, commonly 5%, and compute a test statistic based on the sample. If the observed value falls in the critical region, we reject H₀.
我们选择一个显著性水平，通常为5%，并基于样本计算检验统计量。如果观察值落在临界区域内，则拒绝H₀。
based on the observed total in a fixed time interval, following a Poisson or normal model.
例如，系统管理员可能检验每小时平均错误数是否从λ = 2增加。检验统计量基于固定时间间隔内的观察总数，遵循泊松模型或正态模型。
In S2, you may be asked to set up H₀ and H₁, calculate the probability of a result as extreme as the observed one, and state conclusions in the context of the problem.
在S2中，您可能会被要求建立H₀和H₁，计算观察结果或更极端结果的概率，并结合问题情境说明结论。
10. Confidence Intervals | 置信区间
alues for an unknown population mean μ. For a known variance σ² and a large sample, the 95% confidence interval for the sample mean x̄ is x̄ ± 1.96 · σ/√n.
置信区间给出未知总体均值μ的一个合理取值区间。当方差σ²已知且样本较大时，样本均值x̄的95%置信区间为x̄ ± 1.96 · σ/√n。
x̄ ± z(α/2) · σ/√n
The width of the interval depends on the confidence level, the population standard deviation, and the sample size. Doubling the sample size reduces the margin of error by a factor of √2.
区间宽度取决于置信水平、总体标准差和样本量。将样本量加倍可使误差幅度减小约√2倍。
In software testing, confidence intervals can be used to estimate the average time to complete a task, giving a range rather than a single guess.
在软件测试中，置信区间可用于估计完成任务的平均时间，给出一个范围而不是单一猜测值。
11. Exam Strategies for S2 | S2考试策略
tandardising formula Z = (X − μ)/σ, and the variance formula are the most frequently used.
首先写下关键公式。泊松公式、标准化公式Z = (X − μ)/σ以及方差公式是最常使用的。
When working with continuous distributions, always check that the total area under the probability density function is 1, and carefully write limits on integrals and inequalities.
在处理连续分布时，务必检查概率密度函数下总面积为1，并仔细写出积分限和不等号方向。
For hypothesis tests, state H₀ and H₁ in words too, not just symbols. Use the correct tail of the distribution (one-tailed or two-tailed) and compare the test statistic with the critical value.
对于假设检验，除了符号外，还要用文字说明H₀和H₁。正确使用分布的单尾或双尾，并将检验统计量与临界值进行比较。
Finally, practise with past AQA S2 papers and interpret every numerical answer in context. Examiners reward clear reasoning and correct notation.
最后，用历年AQA S2试卷练习，并在实际情境中解释每个数值答案。考官重视清晰的推理和正确的符号表达。
Published by TutorHao | Computer Science Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图