Week 7 + 8: Machine Learning

Question 1

What is machine learning

Answer 1

learning from data without previous programming used to discover hidden patterns/trends enables data driven decisions

Question 2

4 categories of machine learning

Answer 2

1. Classification
2. Regression
3. Clustering
4. Association analysis

Question 3

Classification is used to predict a…

Answer 3

category

Question 4

Regression is used to predict a…

Answer 4

numeric value

Question 5

Cluster analysis is used to

Answer 5

organise simliar items into groups eg customers

Question 6

Association analysis is used to…

Answer 6

capture assocations between items or events

Question 7

Name 2 supervised machine learning techniques

Answer 7

1. Classification
2. Regression

Question 8

Name 2 unsupervised machine learning techniques

Answer 8

1. Clustering
2. Association analysis

Question 9

What is supervised machine learning

Answer 9

Where you have input variables an output variable using an algorithm to learn the mapping function between the input and the output

Question 10

4 examples of supervised machine learning algorithms

Answer 10

1. KNN
2. Decision tree
3. Linear Regression
4. SVM (Support vector machines)

Question 11

What is unsupervised machine learning? It is where you only have…

Answer 11

input data and not corresponding output variables

Question 12

What is the goal of unsupervised machine learning?

Answer 12

To model the underlying structure or distribution in the data

Question 13

2 examples of unsupervised machine learning algorithma

Answer 13

1. k-means clustering
2. apriori for association analysis

Question 14

kNN is used to

Answer 14

classify a sample based on its neighbors

Question 15

What is k in kNN? The value of k determines…

Answer 15

the number of nearest neighbors to consider

Question 16

kNN 4 distance metrics

Answer 16

1. Euclidean Distance
2. City Block Distance
3. Chi square distance
4. Cosine distance

Question 17

2 pros of kNN

Answer 17

1. No separate training phase
2. Can generate complex decision boundaries

Question 18

2 cons of kNN

Answer 18

1. Can be susceptible to noise
2. Can be slow, since distance is recalculated each time

Question 19

What is decision tree ? It is a..

Answer 19

hierarchical structure with nodes and directed edges

Question 20

Name 3 parts of decision tree

Answer 20

1. Root node - node at the top
2. Internal nodes - in between
3. Leaf nodes - nodes at the bottom

Question 21

A decision tree classification decision is made by

Answer 21

• traversing the decision tree from the root node
• answer to the test condition determines the branch when leaf node is reached
• the category at the leaf node is the classifiction

Question 22

What is decision tree - depth of a node

Answer 22

number of edges from the root node to that node

Question 23

What is decision tree - depth of a decision tree

Answer 23

number of edges in the longest path from the root node to the leaf node

Question 24

What is decision tree - size of a decision tree

Answer 24

the number of nodes in the tree

Question 25

When to stop splitting a node?

Answer 25

• All samples in the node have the same class label
• Max tree depth is reached
• change in impurity is reached

Question 26

2 pros of decision tree

Answer 26

• resulting tree is easy to interpret induction is computationally inexpensive

Question 27

2 cons of decision tree

Answer 27

• greedy approach does not guarantee best solution
• rectilinear decision boundaries

Question 28

What is linear regression?

Answer 28

a statistical method that allows us to summarise the relationship between two continuous variables

Question 29

In linear regression what 3 names can the X variable be called?

Answer 29

1. predictor
2. explanatory
3. independent variable

Question 30

In linear regression what 3 names can the Y variable be called?

Answer 30

1. response
2. outcome
3. dependent variable

Question 31

In one dimension a hyperplane is called a

Answer 31

point

Question 32

In two dimensions a hyperplane is called a

Answer 32

line

Question 33

In one dimensions a hyperplane is called a

Answer 33

plane

Question 34

In 4 or more dimensions a hyperplane is called a

Answer 34

hyperplane

Question 35

The goal of a SVM is to find the…

Answer 35

optimal separating hyperplane which maximizes the margin of the training data

Question 36

3 pros of SVM

Answer 36

1. works well with a clear margin of separation
2. effetive in high dimensional spaces
3. effective in cases where number of dimensions is greater than the number of samples

Question 37

3 cons of SVM

Answer 37

1. doesn’t work well with large data set
2. doesn’t perform very well with noisy data
3. doesn’t provide probability estimates

Question 38

feed forward NN indicates there are

Answer 38

no loops in the network

Question 39

feedback NN

Answer 39

is also known as a recurrent neural network

Question 40

4 pros of NN

Answer 40

• can be trained directly on data with thousands of input variables
• once trained predictions are fast
• good for complex problems (image recognition)
• out-performs other models with high quality labelled data

Question 41

4 cons of NN

Answer 41

• black box
• training is computationally expensive
• suffers from interference where new data causes to forget old data
• often abused where simpler solutions such as linear regression would be best

Question 42

What distance calucation does K means clustering use?

Answer 42

Euclidean distance

Question 43

What does the K in K-Means represent?

Answer 43

The amount of clusters to divide into

Question 44

What type of data does K-means clustering work with?

Answer 44

Continuous

Question 45

What does K-means cluster try to impove?

Answer 45

The inter group simliarity while keeping the groups as far as possible from each other