Scribd
Upload
11 views2 pages

Exercise Sheet 2

1. The ML classifier achieved a precision of 30/35=86%, recall of 30/35=86%, and accuracy of 130/150=87% on the validation set based on the given confusion matrix. 2. Exercises 2-4 involve performing normalization, standardization, and outlier detection on students' exam score data. 3. Exercises 5-6 cover properties of independent and uncorrelated random variables, namely that independence implies uncorrelation, and the definition of uncorrelation in terms of expected values. 4. Exercise 7 involves determining the covariance and correlation coefficient between two random variables where one is a linear transformation of the other.

Uploaded by

nathangame465
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
11 views2 pages

Exercise Sheet 2

1. The ML classifier achieved a precision of 30/35=86%, recall of 30/35=86%, and accuracy of 130/150=87% on the validation set based on the given confusion matrix. 2. Exercises 2-4 involve performing normalization, standardization, and outlier detection on students' exam score data. 3. Exercises 5-6 cover properties of independent and uncorrelated random variables, namely that independence implies uncorrelation, and the definition of uncorrelation in terms of expected values. 4. Exercise 7 involves determining the covariance and correlation coefficient between two random variables where one is a linear transformation of the other.

Uploaded by

nathangame465
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Exercise sheet: End-to-end ML project

The following exercises have different levels of difficulty indicated by (*), (**), (***). An exercise with (*)
is a simple exercise requiring less time to solve compared to an exercise with (***), which is a more complex
exercise.

1. (*) You have built an ML classifier that detects whether a tissue appearing in an image is cancerous
or not. Consider the cancerous class as the positive class. The following confusion matrix shows the
predicted results obtained in the validation set

cancerous (predicted) healthy (predicted)


cancerous (actual) 30 5
healthy (actual) 15 100

Compute the precision, recall and accuracy of your ML classifier.

2. (*) Table 1 below shows the scores achieved by a group of students on an exam. Using this data,
perform the following tasks on the Score feature
(a) A normalisation in the range [0, 1].
(b) A normalisation in the range [−1, 1].
(c) A standardisation of the data.

ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Score 42 47 59 27 84 49 72 43 73 59 58 82 50 79 89 75 70 59 67 35

Table 1: Students’ score

3. (*) We designed a model for predicting the number of bike rentals (y ) from two attributes, temperature
(x1 ) and humidity (x2 ),

y = 500 × x1 + 300 × x2 .

The model was trained with normalised data with values min x1 = −10 and max x1 = 39 for x1 , and
values min x2 = 20 and max x2 = 100. At test time, the model is used to predict the bike rentals for a
vector x∗ = [25, 70]⊤ . What is the value of the prediction y ?

4. (*) A simple criterion to remove outliers from a dataset is to compute the mean, µ, and the standard
deviation, σ , of the variable of interest and consider values outside the range (µ − 3σ, µ +3σ ) as outliers.
Applying this criterion to the Scores in Exercise 2, which ones of them can be considered as outliers?

5. (**) Suppose the joint pmf of the two RVs X and Y is given as
(
1
, for (x1 = 0, y1 = 1), (x2 = 1, y2 = 0), (x3 = 2, y1 = 1)
P (X = xi , Y = yj ) = 3
0 otherwise,

1
(a) Are X and Y independent?
(b) Are X and Y uncorrelated?

6. (**) Two RVs X and Y are uncorrelated if σX,Y = 0. Since σX,Y = E {X Y } − E {X }E {Y }, the two
RVs are uncorrelated if E {X Y } = E {X }E {Y }. Show that if the RVs are independent, then they are
also
 uncorrelated.
HINT: the expected value E {X Y } is defined as
XX
E {X Y } = xi yj P (xi , yj ),
∀xi ∀yj

where P (xi , yj ) is the joint pmf for the discrete RVs X and Y .  A similar definition can be written if
X and Y are continuous RVs, replacing the sums for integrals.

7. (***) Let Y = aX + b, where Y and X are RVs and a and b are constants.

(a) Find the covariance of X and Y .


(b) Find the correlation coefficient of X and Y .

You might also like