Chapter 2

Question 1

Data science is a concept used to tackle big data and includes data cleansing, preparation, processing, and ______.

Answer 1

analysis

Question 2

A data scientist gathers data from multiple sources and applies ______ ______ to extract critical information from the collected data sets.

Answer 2

machine learning

Question 3

The key idea is to convert real-world problem into well-defined ______ ______ problem, so that it can be solved using machine learning.

Answer 3

data science

Question 4

Data mining is the process of analyzing vast amounts of data from various sources to extract ______ ______.

Answer 4

useful information

Question 5

Data mining is done through the discovery of previously unknown patterns, correlations, and ______, which can then be used to predict future outcomes.

Answer 5

anomalies

Question 6

In binary classification, there are only ______ possible classes (or labels) for each instance.

Answer 6

two

Question 7

An example of binary classification is ______ ______ where an email is either spam or not spam.

Answer 7

Spam detection

Question 8

In multiclass classification, there are more than ______ possible classes for each instance.

Answer 8

two

Question 9

An example of multiclass classification is ______ ______ where handwritten digits (0-9) need to be classified into one of the ten classes.

Answer 9

Digit recognition

Question 10

The key difference between binary and multiclass classification regarding the number of classes is that binary has 2 classes, while multiclass has ______ or more.

Answer 10

3

Question 11

In multiclass classification, the output is a ______ of ______, with each probability corresponding to a different class.

Answer 11

vector,probabilities

Question 12

Supervised learning discovers patterns in the data that relate data attributes with a ______ (______) attribute.

Answer 12

target,class

Question 13

In supervised learning, each data point in the training set has an associated ______ or ______.

Answer 13

label,output

Question 14

Two examples of tasks in supervised learning are ______ and ______.

Answer 14

Classification,Regression

Question 15

Unsupervised learning involves training a model on a dataset without any ______ ______.

Answer 15

labeled responses

Question 16

The goal of unsupervised learning is to discover underlying ______, ______, or ______ in the data.

Answer 16

patterns,structures,relationships

Question 17

Two examples of tasks in unsupervised learning are ______ and ______ ______.

Answer 17

Clustering,Dimensionality Reduction

Question 18

Supervised learning has a ______ mechanism, while unsupervised learning does not.

Answer 18

feedback

Question 19

Supervised learning is used for ______, while unsupervised learning is used for ______.

Answer 19

prediction,analysis

Question 20

Algorithms for supervised learning include decision trees, logistic regressions, and ______ ______ ______.

Answer 20

support vector machine

Question 21

Algorithms for unsupervised learning include k-means clustering, hierarchical clustering, and ______ ______.

Answer 21

apriori algorithm

Question 22

A ______ Problem in data science asks ‘How much? How many?’

Answer 22

Regression

Question 23

A ______ Problem in data science asks ‘Is it type A or type B or type C?’

Answer 23

Classification

Question 24

A ______ Problem in data science asks ‘How is the data organized?’

Answer 24

Clustering

Question 25

An ______ ______ Problem in data science asks 'Is it a weird behavior?'

Answer 25

Anomaly Detection

Question 26

A ______-______ and ______ Problem in data science asks 'What should be done next?'

Answer 26

Decision-making,Planning

Question 27

In a regression problem like predicting game sales, the model could be represented as: Sales = f (______).

Answer 27

Variables

Question 28

In a classification problem like digit recognition, the model tries to find the Probability of being a digit in terms of other Variables, like ______ ______.

Answer 28

pixel values

Question 29

In a decision-making and planning problem, the aim is to model a ______/______ Function and try to Maximize/Minimize it.

Answer 29

Profit/Loss

Question 30

Clustering groups data instances that are ______ to (near) each other in one cluster and data instances that are very ______ (far away) from each other into different clusters.

Answer 30

similar,different

Question 31

Clustering is often called an ______ ______ task as no class values denoting an a priori grouping of the data instances are given.

Answer 31

unsupervised learning

Question 32

Association rule mining is also considered an ______ learning task.

Answer 32

unsupervised

Question 33

An example of clustering in real-life is grouping people of similar sizes together to make 'small', 'medium' and 'large' ______.

Answer 33

T-Shirts

Question 34

In marketing, clustering is used to ______ customers according to their similarities to do targeted marketing.

Answer 34

segment

Question 35

Two main types of clustering algorithms are ______ clustering and ______ clustering.

Answer 35

Partitional,Hierarchical

Question 36

Clustering quality is assessed by maximizing ______-______ distance and minimizing ______-______ distance.

Answer 36

Inter-clusters,Intra-clusters

Question 37

Anomaly detection problems try to find ______ of the Data compared to the Regular Pattern observed through the data model.

Answer 37

Deviations

Question 38

Outliers are the set of data points that are ______ ______ than the remainder of the data.

Answer 38

considerably different

Question 39

Bad data quality due to anomalies can lead to bad statistical tests, dashboards, machine learning models, and ultimately, a compromised foundation for ______ ______-______.

Answer 39

informed decision-making

Question 40

The five main types of data science problems are well-posed using T (Type of data science problem), P (How well the problem is answered), and E (______ ______).

Answer 40

Available Data

Question 41

A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, ______ with experience E.

Answer 41

improves

Question 42

To solve a real-world problem using machine learning, we need to identify: T (task), P (______ ______), and E (training experience).

Answer 42

performance measure

Question 43

For a handwriting recognition learning problem, the Task T is recognizing handwritten words, Performance measure P could be the ______ of words correctly classified, and Training experience E would be a ______ of handwritten words with known classifications.

Answer 43

percentage,database

Question 44

For a robot driving learning problem, the Task T is driving on public highways using vision sensors, Performance measure P could be the average ______ traveled before an error, and Training experience E would be a sequence of ______ and ______ commands recorded while observing a human driver.

Answer 44

distance,images,steering

Question 45

To quote the 'optimal' sale price for a car using machine learning, this is a ______ Problem.

Answer 45

Regression

Question 46

To determine how many types of flowers are there in Malaysia using machine learning, this is a ______ Problem.

Answer 46

Clustering

Question 47

Scikit-Learn implements many popular machine learning methods, data processing ______ and visualizations.

Answer 47

pipeline

Question 48

TensorFlow is an open-source ______ ______ Library.

Answer 48

Deep Learning

Question 49

A well-defined learning problem requires a well-specified task, ______ ______, and source of ______ ______.

Answer 49

performance metric,training experience