Topic 5: Machine Learning: Classification & Clustering

Question 1

Calculate the Euclidean distance

Answer 1

SQRT((X_b-X_a)^{2<span> </span>}+ (Y_b-Y_a)²)

Question 2

Define nearest neighbors and combining function.

Answer 2

Nearest neighbours are the most similar instances; where combining function will give us a prediction (through voting/averaging).

Question 3

Explain how combining functions can be used for classification.

Answer 3

looking at the nearest neighbours and using a combining function, such as majority vote, to determine which class the instance belongs to.

Question 4

Calculate the probability of belonging to a class based on nearest
neighbor classification.

Answer 4

number of confirmations of belonging to that instance / total number of k instances

Question 5

Explain weighted voting (scoring) or similarity moderated voting (scoring)

Answer 5

Weighted scoring: influence of neighbours drops the further they’re from the instance

Similarity moderated voting:

Question 6

Explain how k in k-NN can be used to address overfitting.

Answer 6

1-NN memorizes the training data (very complex model). To address overfitting try different values for k and choose the one that gives the best performance on training data and evaluate this on the test data.

Question 7

Discuss issues with nearest-neighbor methods with a focus on
• Intelligibility
• Dimensionality and domain knowledge
• Computational efficiency.

Answer 7

• Intelligibility - if intelligebility and justification are critical, NN should be avoided
• Dimensionality and domain knowledge - curse of dimensionality (all attributes add to the distance, not all attributes are relevant)
• Computational efficiency - training is very fast, prediction/classification of new instance is very inneficient/costly.

Question 8

Describe feature selection.

Answer 8

Selecting features that should be included in the model, can be done manually by someone with industry knowledge.

Question 9

Define and discuss the curse of dimensionality.

Answer 9

Some features are irrelevant but all of the features add to the distance calculations (misleading and confusion of the model).

Question 10

Calculate the Manhattan distance and the Cosine distance

Answer 10

Manhattan distance = (X_a-X_b)+(Y_a-Y_b)

Question 11

Define the Jaccard distance

Answer 11

overlapping items / total unique items

Question 12

Calculate edit distance or the Levenshtein metric

Answer 12

Number of changes it takes to change one text into another by used three actions:

insert, modify, or delete. It is used when order is important.

CAT to FAT (one modify action)

LD = 1

Question 13

Define clustering, hierarchical clustering, and dendrogram

Answer 13

clustering: unsupervised segmentation

hierarchical clustering: overlap between clusters where one cluster contains other clusters

dendogram: hierarchy of the clusters

Question 14

Describe how a dendrogram can help decide the number of clusters.

Answer 14

Horizontal lines can cut across at any point to get to the desired number of clusters.

Question 15

Describe the advantage of hierarchical clustering.

Answer 15

It allows you to see the groupings (ie landscape of data similarity)

Question 16

Define linkage functions.

Answer 16

Distance functions between instances or clusters.

Question 17

Describe how distance measures can be used to decide the number of clusters
in a dendrogram.

Answer 17

1. choosing the line which yields the most clusters and removes the longest distances
2. very long distances are outliers (usually also its own cluster)

Question 18

Define “cluster center” or centroid and k-means clustering

Answer 18

cluster-center: geometric center of a group of instances

k-means clustering: means are the centroid (arithmetic mean) of the values along each dimension for instances in the cluster.

Question 19

Compare and contrast k-means clustering with hierarchical clustering

Answer 19

k-means starts with a desired number of k clusters

Question 20

Describe the k-means algorithm

Answer 20

1. find the points closest to the chosen centers (often random)
2. find the actual center of the clusters found in the first step

Question 21

Describe the reason for running the k-means algorithm many times.

Answer 21

result of a single clustering will find local optimum dependent on the initial centroid locations.

Question 22

Define a cluster’s distortion

Answer 22

sum of the squared differences between each data point and its corresponding centroid.

Question 23

Describe the method for selecting k in the k-means algorithm.

Answer 23

1. experiment with different k-values
2. graph various versions of K (elbow plot) and select the k where stabilization begins

Question 24

Define and calculate the accuracy and error rate

Answer 24

A general measure of classifier performance

Accuracy = Number of correct decisions / Total Number Of Decisions
(equal to 1-Error rate)

Question 25

Describe a confusion matrix.

Answer 25

Summary of prediction results on a classification problem (n x n matrix)

Question 26

Define false positives and false negatives

Answer 26

false positives (negative instances classified as positives)

Question 27

Describe unbalanced data and the problems with unbalanced data

Answer 27

unbalanced data -> where one class is rare evaluation based on accuracy breaks down

Question 28

Discuss the problems with unequal costs and benefits of errors.

Answer 28

Simple classification accuracy as metric makes no distinction between false positives and false negatives (they are equally important). Ideally you would estimate the cost or benefit for each decision a classifier can make.

Question 29

Calculate expected value and expected benefit.

Answer 29

expected benefit: probability_response(x) * (value response) + [1-probability_response]*(value no response)

Question 30

Describe how expected value can be used to frame classifier use

Answer 30

if the expected value is greater than 0 target the customer

Question 31

Describe how expected value can be used to frame classifier evaluation

Answer 31

you can use expected value to compare models.

Question 32

Define and interpret precision and recall.

Answer 32

precision: TP / (TP + FP) = how many times does the model correctly predict a cancer patients out of the total positive predictions.
recall: TP / FN -> how many times did the model correctly predict cancer patients out of the total number of cancer patients

Question 33

Calculate the value of the F-measure.

Answer 33

(precision * recall) / (precision + recall)