Lecture 1

Question 1

Machine Learning

Answer 1

the field of study that gives computers the ability to learn without being explicitly programmed

Question 2

Supervised Learning

Answer 2

An algorithm maps a new input to an output based on example input-output pairs of the training data.

Question 3

Unsupervised Learning

Answer 3

Only the input data is known, and no known output data is given to the algorithm

Question 4

Accuracy

Answer 4

the fraction of inputs for which the right output was predicted

Question 5

Training Data

Answer 5

Data used to build a machine learning model

Question 6

Test Data

Answer 6

Data used to assess how well the model works

Question 7

How is a z-score computed?

Answer 7

Subtracting the mean and dividing by the standard-deviation

Question 8

Reinforcement Learning

Answer 8

Involves reasoning under uncertainty and how agents take actions to maximize their reward

Question 9

Semi-supervised Learning

Answer 9

Involves a small portion of labeled examples and a large number of unlabeled examples from which a model must learn and make predictions on new examples

Question 10

Active Learning

Answer 10

A learning algorithm can interactively query a user to label new data points with the desired outputs

Question 11

Model

Answer 11

An equation that links the values of some features to the predicted value of the target variable

Question 12

Score functions/Fit statistics/Score metrics

Answer 12

measures of how well

the model fits the data

Question 13

Feature selection

Answer 13

reducing the number of predictors by selecting the important ones (dimensionality reduction)

Question 14

Feature extraction

Answer 14

reducing the number of predictors by means of a 
mathematical operation (e.g., PCA)

Question 15

Model Building

Answer 15

finding the equation of the model and the coefficients in it

Question 16

What are two typical tasks for Machine Learning?

Answer 16

1. Prediction (supervised learning)

2. To learn something previously unknown (unsupervised learning)

Question 17

What are the two main types of Supervised Learning?

Answer 17

Classification and Regression

Question 18

Classification

Answer 18

A discrete output such as color, gender, yes/no, class membership

question example: “Will you pass this course?”

Question 19

Regression

Answer 19

A continuous output like temperature, age, distance, salary

question example: “How many points will you get in the exam?”

Question 20

Preprocessing

Answer 20

Cleaning and/or transforming the data

Question 21

When do machine learning algorithms not preform well?

Answer 21

When the input numerical attributes have a very different scale

Question 22

Standard Scaler

Answer 22

z-scores or standard scores where the mean is 0 and the standard deviation is 1

Question 23

What type of data does the Standard Scaler work well with?

Answer 23

it’s a common method in data normalization so it’s good for non-skewed data

Question 24

Robust Scaler

Answer 24

The median is 0 and the interquartile range is 1

Question 25

What type of data does the Robust Scaler work well with?

Answer 25

It’s better for skewed data because it deals better with outliers

Question 26

MinMax Scaler

Answer 26

Shifts data to an interval set by Xmin and Xmax.

Formula:

Xnew = (x - Xmin) / (Xmax - Xmin)

Question 27

What happens when you log scale your data?

Answer 27

You get a better prediction accuracy

Question 28

Normalizer

Answer 28

Each row of the data is rescaled so that its norm becomes 1. Doesn’t work by feature (column) and is only used when the direction of the data matters

Question 29

What type of graph is a normalizer helpful for

Answer 29

histograms

Question 30

Binning

Answer 30

Separating the feature values into n categories. You can replace all the values within each category with a single value like their mean

Question 31

What is Binning effective for?

Answer 31

Models with few parameters like regression models

Question 32

What is Binning not effective for?

Answer 32

Models with many parameters like decision trees

Question 33

Cross-validation

Answer 33

Evaluates the model’s ability to predict new data; detects overfitting or selection bias

Question 34

Feature

Answer 34

properties that describe data points

Question 35

Sample/Instance

Answer 35

a data point; each entity or row in the data

Question 36

Pipeline

Answer 36

The end-to-end construct that orchestrates the flow of data into, and output from, a machine learning model

Question 37

Clustering

Answer 37

A type of unsupervised learning where the algorithm finds natural groups or clusters in data

Question 38

Feature Vector

Answer 38

A vector listing all the feature values

Question 39

Feature Value

Answer 39

The value of a property or feature of the data point/instance, e.g. white, 66, yes

Question 40

Features

Answer 40

An individual measurable property or characteristic of a data point, e.g. color, age, is rich

Question 41

True or false: Classification problems can be used to predict only two discrete valued output such as 0 and 1.

Answer 41

False

Classification can be used for an arbitrary number of classes, not just 2.

Question 42

Which scaling method results in a range between 0 and 1?

Answer 42

Min-Max scaler

Question 43

The standard scaler uses z-scores. How do you compute z-scores?

Answer 43

Subtracting the mean and dividing by the standard-deviation