Key Mistakes in Data Compression (IB Mathematics: Applications and Interpretation HL) | 数据压缩易错点总结（IB数学应用与解释HL）

📚 Key Mistakes in Data Compression (IB Mathematics: Applications and Interpretation HL) | 数据压缩易错点总结（IB数学应用与解释HL）

Data compression is a fascinating topic in the IB Mathematics: Applications and Interpretation HL syllabus, combining probability, logarithms, and algorithms. Students often lose marks due to subtle misconceptions. This article highlights the most frequent pitfalls and how to avoid them, helping you master the concepts tested in exams.

数据压缩是IB数学应用与解释HL课程中一个引人入胜的主题，融合了概率、对数与算法知识。学生在考试中常因一些细微的误解而失分。本文将重点剖析最常见的错误陷阱及其避免方法，帮助你扎实掌握考点。

1. Confusing the Logarithm Base in Entropy | 混淆熵公式中的对数底数

Entropy in information theory is measured in bits, so the formula H = −∑ p(x) log₂ p(x) must use log base 2. A frequent mistake is to use the natural logarithm (ln) or common logarithm (log₁₀). Using the wrong base gives a value that is not in bits and leads to incorrect comparisons.

信息论中的熵以比特为单位，因此公式 H = −∑ p(x) log₂ p(x) 必须使用以2为底的对数。常见错误是使用自然对数 (ln) 或常用对数 (log₁₀)。使用错误的底数会得到非比特单位的数值，从而导致比较错误。

Always check your calculator: log₂ is usually obtained via the change of base formula log₂(a) = ln(a)/ln(2) or log(a)/log(2). Many exam questions provide all probabilities and expect explicit use of log₂.

始终检查计算器设置：log₂ 通常通过换底公式 log₂(a) = ln(a)/ln(2) 或 log(a)/log(2) 来获得。许多考题直接给出概率，期望考生明确使用 log₂ 计算。

2. Forgetting the Negative Sign in Entropy | 遗漏熵公式中的负号

The entropy formula is H = −∑ p(x) log₂ p(x). Since probabilities are between 0 and 1, log₂ p(x) is negative, so the negative sign makes the sum positive. Missing the minus sign yields a negative entropy, which is impossible. Students often write the sum without the minus and then misinterpret the result.

熵的公式为 H = −∑ p(x) log₂ p(x)。由于概率值在0到1之间，log₂ p(x) 为负数，所以负号使得求和结果为正。遗漏负号会得到负的“熵”，这在物理上是不可能的。学生常不加负号直接求和，然后对结果产生困惑。

When calculating by hand, record each term carefully: for p = 0.2, the contribution is −0.2 × log₂(0.2) which is positive. Always verify that your final answer is positive.

手工计算时，仔细记录每一项：对于 p = 0.2，贡献为 −0.2 × log₂(0.2)，这是一个正数。务必确认最终答案为正。

3. Maximum Entropy and Equal Probabilities Misunderstanding | 最大熵与等概率误解

Entropy is maximised when all outcomes are equally likely. For an alphabet of N symbols, the maximum entropy is H_max = log₂ N bits. A common error is to think that any symmetric-looking distribution gives maximum entropy, or to use N = number of trials instead of number of distinct symbols.

当所有结果等可能时，熵达到最大值。对于包含 N 个符号的字母表，最大熵为 H_max = log₂ N 比特。常见错误是认为任何看起来对称的分布都能给出最大熵，或将 N 误认为是试验次数而非不同符号的个数。

For example, with probabilities {0.5, 0.5}, H_max = log₂ 2 = 1 bit, which matches the actual entropy. With {0.7, 0.3}, the entropy is lower. Never confuse the number of categories with sample size.

例如，对于概率 {0.5, 0.5}，H_max = log₂ 2 = 1 比特，与实际熵一致。对于 {0.7, 0.3}，熵会更低。切勿混淆类别数量与样本大小。

4. Huffman Coding – Incorrect Merging Order | 哈夫曼编码——合并顺序错误

When constructing a Huffman tree, you repeatedly combine the two symbols with the smallest probabilities. A frequent slip is to choose the wrong pair when probabilities are equal, or to forget to reinsert the new combined node into the list with its total probability. This results in a non-optimal code and a longer average codeword length.

构建哈夫曼树时，需要反复合并概率最小的两个符号。常见的失误是在概率相等时选错了合并对象，或忘记将合并后的新节点及其总概率重新放回列表。这会导致编码非最优，平均码字长度变长。

Always list all nodes (including combined ones) in ascending order of probability before each step. If two equal smallest probabilities can be chosen, either choice is valid, but you must stick to a consistent tie-breaking rule (e.g., leftmost smallest).

每一步操作前，始终将所有节点（包括已合并的）按概率升序排列。如果存在两个相等的极小概率，任选其一均有效，但必须保持一致的平局处理规则（例如总是选择最靠左侧的最小值）。

5. Average Codeword Length Calculation Errors | 平均码字长度计算错误

The average length is L = ∑ p(x) × length(x). Mistakenly using the reciprocal of probability as length, or summing lengths without weighting, are typical blunders. Some students also forget that lengths are the number of bits used by each codeword, not the size of the symbol itself.

平均长度为 L = ∑ p(x) × length(x)。典型错误包括误用概率的倒数作为长度，或不进行加权直接累加长度。有些学生还忘记长度是各码字使用的比特数，而非符号自身的大小。

After building the Huffman tree, carefully trace the depth of each leaf to obtain its codeword length. Double-check that you haven’t swapped probability and length in the sum.

构建哈夫曼树后，仔细追溯每个叶节点的深度以获得其码字长度。务必检查在求和时没有把概率和长度的位置弄反。

6. Compression Ratio vs. Space Savings Confusion | 压缩比与空间节省混淆

The compression ratio is usually defined as (original size) / (compressed size). A high compression ratio means greater reduction. Space saving is (1 − compressed/original) × 100%. Students often invert the ratio or mislabel the percentage, leading to completely wrong conclusions about efficiency.

压缩比通常定义为（原始大小）/（压缩后大小）。压缩比越高意味着缩减幅度越大。空间节省为 (1 − 压缩后/原始) × 100%。学生常颠倒比值或错误标注百分比，导致关于效率的结论完全错误。

When a problem gives the number of bits before and after compression, compute clearly: if original uses 800 bits and compressed uses 200 bits, ratio = 4:1, saving = 75%. Keep the definitions straight.

当题目给出压缩前后的比特数时，计算要清晰：若原始使用800比特，压缩后使用200比特，则压缩比为 4:1，节省空间 75%。定义必须清晰。

7. Redundancy and Efficiency Miscalculation | 冗余度与效率计算错误

Redundancy is the difference between average codeword length and entropy: R = L − H bits per symbol. Efficiency is H/L. A common slip is to calculate redundancy as a proportion without checking units, or to think that L can be less than H – it cannot, by the source coding theorem.

冗余度是平均码字长度与熵的差：R = L − H（每符号比特数）。效率为 H/L。常见失误是不检查单位就将冗余度当成比例来计算，或认为 L 可以小于 H——根据信源编码定理，这是不可能的。

If your calculated L is smaller than H, you have likely made an arithmetic mistake or used the wrong logarithm base. Always compare them and ensure L ≥ H.

如果计算出的 L 小于 H，很可能出现了计算错误或对数底数不对。务必对比两者，确保 L ≥ H。

8. Prefix Code Verification and Kraft Inequality Mistakes | 前缀码验证与克拉夫特不等式错误

A code is uniquely decodable if it is a prefix code – no codeword is a prefix of another. To check, you must test every codeword against the others. The Kraft inequality ∑ 2^(−l_i) ≤ 1 gives a necessary and sufficient condition for the existence of a prefix code with lengths l_i. Students often forget to raise 2 to the negative length, or sum probabilities instead.

如果一个码是前缀码——即没有码字是其他码字的前缀，它就是唯一可解码的。验证时必须逐一检查每个码字。克拉夫特不等式 ∑ 2^(−l_i) ≤ 1 为给定长度 l_i 的前缀码存在性提供了充要条件。学生常忘记对2取负长度次幂，或误将概率相加。

When verifying Kraft, compute 2^(−l_i) for each codeword and sum them. The sum must be ≤ 1. This is not the same as summing probabilities. Also, remember that satisfaction of Kraft guarantees the existence of some prefix code, not necessarily the one you have in hand.

验证克拉夫特不等式时，对每个码字计算 2^(−l_i) 并求和，总和必须 ≤ 1。这与概率求和不同。此外，满足克拉夫特不等式仅保证存在某个前缀码，并不一定就是你手中的编码。

9. Misunderstanding Information Content of an Event | 误解事件的信息量

The information content (or self-information) of an event with probability p is I = −log₂ p bits. This measures the surprise of that specific event. A rare event carries more information. A common error is to treat information content as the same as entropy, or to calculate it without the negative sign, or to add probabilities when combining independent events instead of multiplying.

一个概率为 p 的事件的信息量（自信息）为 I = −log₂ p 比特。它衡量该特定事件带来的意外程度。罕见事件携带更多信息。常见错误包括混淆信息量与熵，或不加负号直接计算，或在组合独立事件时对概率进行相加而非相乘。

For multiple independent events, total information is the sum of their individual information contents because I(A and B) = −log₂(pA × pB) = −log₂(pA) − log₂(pB). Ensure you multiply probabilities, not add them.

对于多个独立事件，总信息量等于各自信息量的和，因为 I(A 且 B) = −log₂(pA × pB) = −log₂(pA) − log₂(pB)。注意概率应当相乘，而非相加。

10. Lossy vs Lossless Compression and Entropy’s Limitations | 有损与无损压缩及熵的局限性

Huffman coding and entropy provide a theoretical bound for lossless compression of independent symbols. A key mistake is applying this directly to image or audio files without considering correlations between pixels or samples. Lossy compression (like JPEG) discards some information to achieve higher compression; entropy cannot describe its performance in a simple formula.

哈夫曼编码和熵为独立符号的无损压缩提供了理论上限。关键错误是直接将其应用于图像或音频文件，而未考虑像素或采样点间的相关性。有损压缩（如 JPEG）通过丢弃部分信息来获得更高压缩率；熵无法用一个简单公式描述其性能。

In an image, adjacent pixels are often similar, so the actual entropy of the source is lower than the symbol-wise entropy if you treat each pixel value independently. Understanding this distinction prevents overestimating the minimum possible file size.

在图像中，相邻像素通常相似，因此信源的实际熵如果考虑了这种相关性，会比仅独立处理每个像素值的符号熵更低。理解这一区别可以防止对文件最小可能大小的高估。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）