BUSS6002 W7 Feature Engineering and NLP
BUSS6002 W7 特征工程和 NLP

Prepared by Ella Luo 编写人：Ella Luo

Part 1. Feature Engineering
第 1 部分.特征工程

1. What is feature engineering
   什么是功能工程

• process of constructing predictors from raw data, based on domain knowledge or algorithm requirements
  根据领域知识或算法要求，从原始数据中构建预测器的过程
• Good feature engineering can significantly improve model performance
  良好的特征工程可大幅提高模型性能
• Example: polynomial regression
  示例：多项式回归

2. log Transformation 对数变换

Original Linear regression: AmountSpent $={\beta }_0+{\beta }_1\times$$={\beta }_0+{\beta }_1\times$=beta_(0)+beta_(1)xx=\beta_{0}+\beta_{1} \times Salary $+\epsilon$$+\epsilon$+epsi+\varepsilon
原始线性回归：支出金额 $={\beta }_0+{\beta }_1\times$$={\beta }_0+{\beta }_1\times$=beta_(0)+beta_(1)xx=\beta_{0}+\beta_{1} \times 工资 $+\epsilon$$+\epsilon$+epsi+\varepsilon

• We can see patterns in the residual plots, which indicate non-linear relationship
  我们可以从残差图中看出一些规律，这表明两者之间存在非线性关系
• Also we can see heteroscedasticity in the variance of errors
  我们还可以看到误差方差的异方差性

Log transformation: $\log ($$\log ($log(\log ( Amount Spent $)={\beta }_0+{\beta }_1\times \log ($$)={\beta }_0+{\beta }_1\times \log ($)=beta_(0)+beta_(1)xx log()=\beta_{0}+\beta_{1} \times \log ( Salary $)+\epsilon$$)+\epsilon$)+epsi)+\varepsilon
对数变换： $\log ($$\log ($log(\log ( 支出金额 $)={\beta }_0+{\beta }_1\times \log ($$)={\beta }_0+{\beta }_1\times \log ($)=beta_(0)+beta_(1)xx log()=\beta_{0}+\beta_{1} \times \log ( 工资 $)+\epsilon$$)+\epsilon$)+epsi)+\varepsilon

• The residual plot shows no pattern
  残差图没有显示模式
• Variance is almost constant now
  现在差异几乎保持不变
• Log transformation can help:
  对数变换可以提供帮助：
• Remove certain type of non-linearity
  消除某些类型的非线性
• Stabilise variance 稳定差异

Retransformation bias: 再转换偏差：

• $$$$quad\quad The model after log-transform: $\log (y)={\mathbf{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}+\epsilon$$\log (y)={\mathbf{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}+\epsilon$log(y)=x^(TT) widehat(beta)+epsi\log (y)=\mathbf{x}^{\top} \widehat{\boldsymbol{\beta}}+\varepsilon
  $$$$quad\quad 对数变换后的模型： $\log (y)={\mathbf{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}+\epsilon$$\log (y)={\mathbf{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}+\epsilon$log(y)=x^(TT) widehat(beta)+epsi\log (y)=\mathbf{x}^{\top} \widehat{\boldsymbol{\beta}}+\varepsilon
• $$$$quad\quad Thus, if we need to predict $\hat{y}$$\widehat{y}$hat(y)\hat{y} :
  $$$$quad\quad 因此，如果我们需要预测 $\hat{y}$$\widehat{y}$hat(y)\hat{y} ：
• $$$$quad\quad Naïve Prediction: ${\hat{y}}_{\text{naive}}=\exp (\widehat{\log (y)})=\exp \left({\mathbf{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}\right)$${\widehat{y}}_{\text{naive }}=\exp (\widehat{\log (y)})=\exp \left({\mathbf{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}\right)$hat(y)_("naive ")=exp( widehat(log(y)))=exp(x^(TT)( widehat(beta)))\hat{y}_{\text {naive }}=\exp (\widehat{\log (y)})=\exp \left(\mathbf{x}^{\top} \widehat{\boldsymbol{\beta}}\right)
  $$$$quad\quad 简单预测： ${\hat{y}}_{\text{naive}}=\exp (\widehat{\log (y)})=\exp \left({\mathbf{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}\right)$${\widehat{y}}_{\text{naive }}=\exp (\widehat{\log (y)})=\exp \left({\mathbf{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}\right)$hat(y)_("naive ")=exp( widehat(log(y)))=exp(x^(TT)( widehat(beta)))\hat{y}_{\text {naive }}=\exp (\widehat{\log (y)})=\exp \left(\mathbf{x}^{\top} \widehat{\boldsymbol{\beta}}\right)
  

  

• Optimal Prediction: ${\hat{y}}_{\text{optimal}}=\mathbb{E}(y\mid \mathbf{x})=\exp \left({\mathit{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}\right)\mathbb{E}[\exp (\epsilon )]={\hat{y}}_{\text{naive}}\times \mathbb{E}[\exp (\epsilon )]$${\widehat{y}}_{\text{optimal }}=\mathbb{E}(y\mid \mathbf{x})=\exp \left({\mathit{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}\right)\mathbb{E}[\exp (\epsilon )]={\widehat{y}}_{\text{naive }}\times \mathbb{E}[\exp (\epsilon )]$hat(y)_("optimal ")=E(y∣x)=exp(x^(TT)( widehat(beta)))E[exp(epsi)]= hat(y)_("naive ")xxE[exp(epsi)]\hat{y}_{\text {optimal }}=\mathbb{E}(y \mid \mathbf{x})=\exp \left(\boldsymbol{x}^{\top} \widehat{\boldsymbol{\beta}}\right) \mathbb{E}[\exp (\varepsilon)]=\hat{y}_{\text {naive }} \times \mathbb{E}[\exp (\varepsilon)]
  最佳预测： ${\hat{y}}_{\text{optimal}}=\mathbb{E}(y\mid \mathbf{x})=\exp \left({\mathit{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}\right)\mathbb{E}[\exp (\epsilon )]={\hat{y}}_{\text{naive}}\times \mathbb{E}[\exp (\epsilon )]$${\widehat{y}}_{\text{optimal }}=\mathbb{E}(y\mid \mathbf{x})=\exp \left({\mathit{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}\right)\mathbb{E}[\exp (\epsilon )]={\widehat{y}}_{\text{naive }}\times \mathbb{E}[\exp (\epsilon )]$hat(y)_("optimal ")=E(y∣x)=exp(x^(TT)( widehat(beta)))E[exp(epsi)]= hat(y)_("naive ")xxE[exp(epsi)]\hat{y}_{\text {optimal }}=\mathbb{E}(y \mid \mathbf{x})=\exp \left(\boldsymbol{x}^{\top} \widehat{\boldsymbol{\beta}}\right) \mathbb{E}[\exp (\varepsilon)]=\hat{y}_{\text {naive }} \times \mathbb{E}[\exp (\varepsilon)]
• According to Jensen’s inequality: $\mathbb{E}[\exp (\epsilon )]>\exp [\mathbb{E}(\epsilon )]=1$$\mathbb{E}[\exp (\epsilon )]>\exp [\mathbb{E}(\epsilon )]=1$E[exp(epsi)] > exp[E(epsi)]=1\mathbb{E}[\exp (\varepsilon)]>\exp [\mathbb{E}(\varepsilon)]=1
  根据詹森不等式： $\mathbb{E}[\exp (\epsilon )]>\exp [\mathbb{E}(\epsilon )]=1$$\mathbb{E}[\exp (\epsilon )]>\exp [\mathbb{E}(\epsilon )]=1$E[exp(epsi)] > exp[E(epsi)]=1\mathbb{E}[\exp (\varepsilon)]>\exp [\mathbb{E}(\varepsilon)]=1
• Thus, ${\hat{y}}_{\text{naïve}}<{\hat{y}}_{\text{optimal}}$${\widehat{y}}_{\text{naïve }}<{\widehat{y}}_{\text{optimal }}$hat(y)_("naïve ") < hat(y)_("optimal ")\hat{y}_{\text {naïve }}<\hat{y}_{\text {optimal }} 因此， ${\hat{y}}_{\text{naïve}}<{\hat{y}}_{\text{optimal}}$${\widehat{y}}_{\text{naïve }}<{\widehat{y}}_{\text{optimal }}$hat(y)_("naïve ") < hat(y)_("optimal ")\hat{y}_{\text {naïve }}<\hat{y}_{\text {optimal }}
• ${\hat{y}}_{\text{unbiased}}:={\hat{y}}_{\text{naive}}\times \frac{1}{n}\sum _{i=1}^n\exp \left({\epsilon }_i\right)=\exp \left({\mathbf{x}}^{\mathrm{\top }}\hat{\mathit{\beta }}\right)\frac{1}{n}\sum _{i=1}^n\exp \left({\epsilon }_i\right)$${\widehat{y}}_{\text{unbiased }}:={\widehat{y}}_{\text{naive }}\times \frac{1}{n}\sum _{i=1}^n \exp \left({\epsilon }_i\right)=\exp \left({\mathbf{x}}^{\mathrm{\top }}\widehat{\mathit{\beta }}\right)\frac{1}{n}\sum _{i=1}^n \exp \left({\epsilon }_i\right)$hat(y)_("unbiased "):= hat(y)_("naive ")xx(1)/(n)sum_(i=1)^(n)exp(epsi_(i))=exp(x^(TT)( widehat(beta)))(1)/(n)sum_(i=1)^(n)exp(epsi_(i))\hat{y}_{\text {unbiased }}:=\hat{y}_{\text {naive }} \times \frac{1}{n} \sum_{i=1}^{n} \exp \left(\varepsilon_{i}\right)=\exp \left(\mathbf{x}^{\top} \widehat{\boldsymbol{\beta}}\right) \frac{1}{n} \sum_{i=1}^{n} \exp \left(\varepsilon_{i}\right)

3. Dealing with categorical Predictors:
3.处理分类预测因子：

1. Using dummy variables to handle qualitative predictors.
   使用虚拟变量处理定性预测因子。

• $k$$k$quad k\quad k - categories need only $k-1$$k-1$k-1k-1 dummy variables
  $k$$k$quad k\quad k - 类别只需 $k-1$$k-1$k-1k-1 虚拟变量
• Reference or Baseline Category: The one that is not explicitly represent by dummy variable
  参考或基准类别：未用虚拟变量明确表示的类别
• Each dummy is compared with the reference when interpreting
  在解释时，将每个假值与参照值进行比较。

Consider the following regression with dummy variable Female:
考虑以下带有虚拟变量 "女性 "的回归：

• AmountSpent $={\beta }_0+{\beta }_1\times$$={\beta }_0+{\beta }_1\times$=beta_(0)+beta_(1)xx=\beta_{0}+\beta_{1} \times Salary $+{\beta }_2\times$$+{\beta }_2\times$+beta_(2)xx+\beta_{2} \times Children $+{\beta }_3\times$$+{\beta }_3\times$+beta_(3)xx+\beta_{3} \times Catalogs $+{\beta }_4\times$$+{\beta }_4\times$+beta_(4)xx+\beta_{4} \times Female $+\epsilon$$+\epsilon$+epsi+\varepsilon
  支出金额 $={\beta }_0+{\beta }_1\times$$={\beta }_0+{\beta }_1\times$=beta_(0)+beta_(1)xx=\beta_{0}+\beta_{1} \times 工资 $+{\beta }_2\times$$+{\beta }_2\times$+beta_(2)xx+\beta_{2} \times 子女 $+{\beta }_3\times$$+{\beta }_3\times$+beta_(3)xx+\beta_{3} \times 目录 $+{\beta }_4\times$$+{\beta }_4\times$+beta_(4)xx+\beta_{4} \times 女性 $+\epsilon$$+\epsilon$+epsi+\varepsilon
• $$$$quad\quad Female $=\{\begin{array}{ll}1,& \text{if Gender}=\text{'Female'}\\ 0,& \text{if Gender}=\text{'Male'}\end{array}$$=\left\{\begin{array}{l}1,\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\text{ if Gender }=\text{ 'Female' }\\ 0,\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\text{ if Gender }=\text{ 'Male' }\end{array}\right.$={[1","," if Gender "=" 'Female' "],[0","," if Gender "=" 'Male' "]:}= \begin{cases}1, & \text { if Gender }=\text { 'Female' } \\ 0, & \text { if Gender }=\text { 'Male' }\end{cases}
  $$$$quad\quad 女 $=\{\begin{array}{cc}1,& \text{if Gender}=\text{'Female'}\\ 0,& \text{if Gender}=\text{'Male'}\end{array}$$=\left\{\begin{array}{c}1,\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\text{ if Gender }=\text{ 'Female' }\\ 0,\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\hspace{0.25em}\text{ if Gender }=\text{ 'Male' }\end{array}\right.$={[1","," if Gender "=" 'Female' "],[0","," if Gender "=" 'Male' "]:}= \begin{cases}1, & \text { if Gender }=\text { 'Female' } \\ 0, & \text { if Gender }=\text { 'Male' }\end{cases}

This means that: 这意味着

• Interpretation: all else being equal, a female customer spending on average ${\hat{\beta }}_4$${\widehat{\beta }}_4$hat(beta)_(4)\hat{\beta}_{4} more than a male customer
  释义：在其他条件相同的情况下，女性顾客的平均消费额 ${\hat{\beta }}_4$${\widehat{\beta }}_4$hat(beta)_(4)\hat{\beta}_{4} 高于男性顾客

2) Interactions: 2) 互动：

Original regression: 原来的退步：

• Interpretation of ${\beta }_1$${\beta }_1$beta_(1)\beta_{1} : the average spending is increased by ${\hat{\beta }}_1$${\widehat{\beta }}_1$hat(beta)_(1)\hat{\beta}_{1} for each dollar increase in salary, if keep the number of children constant
  ${\beta }_1$${\beta }_1$beta_(1)\beta_{1} 的解释：在子女人数不变的情况下，工资每增加一美元，平均支出就会增加 ${\hat{\beta }}_1$${\widehat{\beta }}_1$hat(beta)_(1)\hat{\beta}_{1} 。

Regression with interaction:
交互回归：

• Interpretation: when salary increased by one-dollar, the average spending is increased by ( ${\hat{\beta }}_1+{\hat{\beta }}_3\times$${\widehat{\beta }}_1+{\widehat{\beta }}_3\times$hat(beta)_(1)+ hat(beta)_(3)xx\hat{\beta}_{1}+\hat{\beta}_{3} \times Children)
  解释：当工资增加一美元时，平均支出增加 ( ${\hat{\beta }}_1+{\hat{\beta }}_3\times$${\widehat{\beta }}_1+{\widehat{\beta }}_3\times$hat(beta)_(1)+ hat(beta)_(3)xx\hat{\beta}_{1}+\hat{\beta}_{3} \times 儿童)

3) Polynomial Regression 3) 多项式回归

Part 2. Natural Language Processing (NLP):
第 2 部分：自然语言处理（NLP自然语言处理（NLP）：

1. What is NLP (Natural Language Processing):
   什么是 NLP（自然语言处理）：

• NLP is automated handling of natural human language like speech or text
  NLP 是对语音或文本等人类自然语言的自动处理
• Make computers to understand all meaning and information in text (Unstructured data)
  让计算机理解文本中的所有含义和信息（非结构化数据）

2. Regular Expressions (Regex) - Not really NLP:
   正则表达式 (Regex) - 并非真正的 NLP：

• Can represent highly complex deterministic queries about the sequence of numbers, characters and special symbols.
  可表示有关数字、字符和特殊符号序列的高度复杂的确定性查询。

3. Structuring Unstructured Data
   构建非结构化数据

• Keywords are called Tokens
  关键词被称为代币
• Useful semantic unit for processing
  有用的语义处理单元
• In English: often identified by whitespace and / or punctuation around it
  在英语中：通常通过周围的空白和/或标点符号来识别
  (1) Bag-of-words (BOW): (1) 词袋（BOW）：

Raw text Bag-of-words representation
原始文本 字袋表示法

• Extract features from text that describes the occurrence of words within a document
  从文本中提取描述文档中词语出现情况的特征
• Create a vocabulary of known words
  创建已知单词词汇表
• A measure of the presence of known words
  衡量是否存在已知词语
• Does not capture the order of word (only capture the occurrence)
  不捕捉词序（只捕捉出现次数）
  (2) Term frequency - inverse document frequency (TF - iDF):
  (2) 词频-反向文档频率（TF-iDF）：

• Translate counts into a measure of how “important” (i.e. unique) a token is to a document relative to a wider collection or corpus.
  将计数转换为衡量一个标记相对于更广泛的集合或语料库对文档的 "重要性"（即唯一性）。

(3) N-gram: (3) N-语法：

• A sequence of $n$$n$nn tokens
  一个 $n$$n$nn 标记序列
• 1-gram: unigram
• 2-gram: bigram
• 3-gram: trigram
• Maintains the structure of the structure, context and possibly the meaning of the original text
  保持原文的结构、上下文以及可能的含义
• Compatible with BoW and TF-iDF (Can be used within BOW and TF-iDF)
  与 BoW 和 TF-iDF 兼容（可在 BOW 和 TF-iDF 中使用）

4. Reduce complexity in unstructured data:
   降低非结构化数据的复杂性：

(1) Stop Words: (1) 停止词：

• Filter out some words that tend to add little to the meaning of a text.
  过滤掉一些对文本意义影响不大的词语。
• Stop words usually refers to the most common words in a language
  停滞词通常指一种语言中最常见的词
• No single universal list of stop words used by all natural language processing tools.
  所有自然语言处理工具都不使用单一的通用停滞词列表。

(2) Removing Grammar with Python's NLTK:
(2) 使用 Python 的 NLTK 删除语法：

• Natural Language Toolkit:
  自然语言工具包
• Stemming: remove the ends of words
  词干：去掉词尾
• Lemmatization: Properly with the use of a vocabulary and morphological analysis of words.
  词法化：正确使用词汇和词形分析。
• Stemming and lemmatization used for removing variations in language
  词干化和词素化用于消除语言中的差异
• This is not statistical or learned algorithm
  这不是统计或学习算法
• Porter 2 (Snowball English Stemmer) is generally the best choice for English
  Porter 2（雪球英语词干转换器）通常是英语学习的最佳选择。

• Sufficiently large training data (the longer of n-gram’s n, the harder to estimate probabilities)
  足够大的训练数据（n-词组的 n 越长，估计概率就越难）
• Many Researchers, use 5-grams or 6-grams
  许多研究人员，使用 5 克或 6 克

5. Sentiment Analysis 情感分析
   (1) VADER (Valence Aware Dictionary and Sentiment Reasoner)
   (1) VADER（价值感知词典和情感推理器）

• A lexicon and rule-based sentiment analysis tool that is specifically to social media (Twitter).
  专门针对社交媒体（Twitter）的基于词典和规则的情感分析工具。
• Assign positive or negative points for each word, then add them up
  为每个单词分配正分或负分，然后将它们相加
• VADER is specially tuned for Twitter
  VADER 专门针对 Twitter 进行了调整
  (2) Word Embedding (2) 单词嵌入
• Problem: frequency representation of words loses a lot of information
  问题：词频表示法丢失了大量信息
• Dog vs poodle 狗与贵宾犬
• Dog vs Cat 狗与猫
• Idea: Represent words as vectors in multidimensional space
  想法将单词表示为多维空间中的向量
• Similar words should be closer
  相似的词语应更接近
• Solution: Word2Vec 解决方案Word2Vec
• Using shallow neural networks to calculate these vectors.
  使用浅层神经网络计算这些向量。

PART 3. Ethics In Data Science
第 3 部分.数据科学的伦理

1. Ethics in the Age of Algorithmic Decisions
1.算法决策时代的伦理问题

• Application of AI (Artificial Intelligence)
  人工智能（AI）的应用
• Not only a technological advance, but also a social change
  不仅是技术进步，也是社会变革
• Data science and Al may harm society to the potential vulnerable populations, misuse personal information and the political manipulations.
  数据科学和人工智能可能会危害社会潜在的弱势群体、滥用个人信息和政治操纵。
• Bias and discrimination are built into AI Algorithms, Al reflects the values of its creators.
  人工智能算法中存在偏见和歧视，这反映了其创造者的价值观。
• Ethical judgement: 道德判断：
• Ethics is concerned with making judgement about intention and actions
  伦理涉及对意图和行动的判断
• Normative theory: norm of society, how things ought to be
  规范理论：社会规范，事物应该如何发展
• Descriptive Theory: explain how things are
  描述理论：解释事物是如何存在的
• Ethical judgements are normative but not descriptive
  道德判断是规范性的，而不是描述性的
• Morality: 道德：
• A system rules that guide human conduct, principles for evaluating rules.
  指导人类行为的规则体系，评价规则的原则。
• Morality is a system 道德是一个体系
• Comprising values and principles. Just like policies.
  包括价值观和原则。就像政策一样。
• Ethics vs Morals 伦理与道德
• Ethics: the rules that a social system provides us with
  道德：社会制度为我们提供的规则
• Morals: Our own principles
  道德：我们自己的原则
• Ethics vs values 道德与价值观

2. Ethics in Data science:
2.数据科学的伦理：

• Data used in training may contain bias
  训练中使用的数据可能存在偏差
• Algorithms themselves learned from the biases and used the conclusions from the biased data to support future decisions.
  算法本身从偏差中学习，并利用从偏差数据中得出的结论来支持未来的决策。
• Usable of data (Data Legality):
  数据的可用性（数据的合法性）：
• Fair Use Agreement: 合理使用协议：
• Harvesting data when a policy in place explicitly forbidding
  在政策明确禁止的情况下采集数据
• Harvesting data behind login page/portal without a policy
  在未制定政策的情况下，在登录页面/门户网站后面采集数据
• Harvesting public data that is not explicitly linked anywhere
  采集未在任何地方明确链接的公共数据
• Harvesting public social media data that is plainly visible might be ethical or unethical or probably legal
  获取公开的社交媒体数据可能是合乎道德的，也可能是不道德的，或者可能是合法的
• UK Gov - 5 principles of ethical AI:
  英国政府 - 人工智能伦理的 5 项原则：
• Should be developed for the common good and benefit of humanity
  应为人类的共同利益和福祉而发展
• Should operate on principles of intelligibility and fairness
  应按照易懂和公平的原则运作
• Should not be used to diminish the data rights or privacy of individuals, families, or communities
  不应削弱个人、家庭或社区的数据权或隐私权
• All citizens should have the right to be educated to enable them to flourish mentally, emotionally, and economically alongside Al
  所有公民都应有权接受教育，使他们能够在精神、情感和经济上与他人并肩发展。
  o The autonomous power to hurt, destroy, or deceive human beings should never be vested in Al.
  o 绝不能将伤害、毁灭或欺骗人类的自主权赋予阿尔。
• Five Common Principles of Responsible AI:
  负责任的人工智能的五项共同原则：

• The field of data ethics offers guiding principles for business professionals who handle data.
  数据伦理领域为处理数据的业务专业人员提供了指导原则。
• Ownership: an individual owns their personal information
  所有权：个人拥有自己的个人信息
• Transparency: Data subjects have a right to know how you plan to collect, store, and use it. When gathering data, exercise transparency.
  透明度：数据主体有权知道您计划如何收集、存储和使用数据。在收集数据时，要做到透明。
• Privacy: Even if a customer gives your company consent to collect, store, and analyse their personally identifiable information (PII), that doesn’t mean that it should be made publicly available.
  隐私：即使客户同意贵公司收集、存储和分析其个人身份信息 (PII)，也并不意味着这些信息应该公开。
• Intention: Before collecting data, ask yourself why you need it, what you’ll gain from it, and what changes you’ll be able to make after analysis. If your intention is to hurt others, profit from your subjects’ weaknesses, or any other malicious goal, it’s not ethical to collect their data.
  意图：在收集数据之前，先问问自己为什么需要这些数据，从中能获得什么，分析后能做出什么改变。如果你的意图是伤害他人、利用研究对象的弱点获利或其他恶意目的，那么收集他们的数据就不道德。
• Outcome: Even when intentions are good, the outcome of data analysis can cause inadvertent harm to individuals or groups of people. This is called a disparate impact, which is outlined in the Civil Rights Act as unlawful.
  结果：即使出发点是好的，数据分析的结果也可能无意中对个人或群体造成伤害。这就是所谓的差异影响，《民权法案》规定这种影响是非法的。

Practice for Week 7. 第 7 周的练习。

1. Developing a classification model for topics in a large corpus of social media posts, you trained a model based on data that represented each document through the count of each token it contains. Dissatisfied with the model’s performance you decide that you need to represent the meaning of the texts more precisely. Which of the following processes would you choose to achieve this?
   在为一个大型社交媒体帖子语料库中的主题开发分类模型时，您训练了一个基于数据的模型，该模型通过包含的每个标记的计数来表示每个文档。由于对模型的性能不满意，您决定需要更精确地表示文本的含义。您会选择以下哪种方法来实现这一目标？
   A. Encode texts as bi-grams.
   A.将文本编码为双语法
   B. Prepare document with stemming.
   B.编写有标点符号的文件。
   C. Remove all stop words.
   C.删除所有停止词。
   D. Engineer features with VADER.
   D.使用 VADER 设计功能。
2. You are tasked to categorise a large set of research papers according to their research disciplines. Which of the following is the most suitable choice to distinguish these documents?
   您的任务是将一大批研究论文按照研究学科分类。以下哪项最适合用来区分这些文件？
   A. TF-iDF A.TF-iDF
   B. VADER B.VADER
   C. Bag-of-Words C.词袋
   D. Regular Expressions (RegEx)
   D.正则表达式 (RegEx)
3. The following regression model is estimated via least squares, using a data-set containing 200 observations:
   利用包含 200 个观测值的数据集，通过最小二乘法估算出以下回归模型：