Scribd
Upload
8 views

Lab1 - Machine Learning Basics (Without Solutions)

The document discusses machine learning basics including defining machine learning, applications, evaluating algorithms, data and plots, learning notation, and evaluating performance. It provides tasks for readers to discuss machine learning concepts like what features and instances in data correspond to, how to plot data, and how to properly evaluate an algorithm's performance on a dataset.

Uploaded by

Sindy Condo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

Lab1 - Machine Learning Basics (Without Solutions)

The document discusses machine learning basics including defining machine learning, applications, evaluating algorithms, data and plots, learning notation, and evaluating performance. It provides tasks for readers to discuss machine learning concepts like what features and instances in data correspond to, how to plot data, and how to properly evaluate an algorithm's performance on a dataset.

Uploaded by

Sindy Condo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Machine Learning Basics

Machine Learning - Lab 1

www.isg.beel.org www.uni-siegen.de
10-15 Minutes
Task 1.1 – What really is Machine Learning?
Discuss (with your neighbors) what Machine Learning could be. Ask yourself:

a) Define Machine Learning in your own words.


b) Is sorting a good application of machine learning?
c) Name 5 applications of machine learning.
d) How can we evaluate machine learning algorithms?
e) What is an optimal function 𝑓𝑓 ∗ in machine learning?
f) What is a model in machine learning? What is a linear model? What is a non-
linear model?
g) What is a prediction in machine learning?

2
10-20 Minutes
Task 1.2 – Data and Plots
Given the data below, discuss (with your neighbors):

a) What could a row (instances 𝑖𝑖1 − 𝑖𝑖5 ) correspond to?


b) What could the columns (features 𝑓𝑓1 , 𝑓𝑓2 ) correspond to?
c) How could you plot the data below? Sketch/draw the data.
d) In your plot, draw the perfect function; draw a good function. Do you see a
problem with the perfect function?
𝒇𝒇𝟏𝟏 𝒇𝒇𝟐𝟐
𝑖𝑖1 1.5 11
𝑖𝑖2 3.0 7
𝑖𝑖3 1.0 12
𝑖𝑖4 2.6 9
𝑖𝑖5 4.0 5 6
10-20 Minutes
Task 1.3 – Learning Notation
Given the data below, transform it into the notation on the right. Discuss what parts of the
notation correspond to what parts of the data. Assume the feature 𝑦𝑦 to be our target.
• Write down 𝒙𝒙𝟏𝟏 ; write down 𝑦𝑦1 ; write down 𝑥𝑥24

𝒇𝒇𝟏𝟏 𝒇𝒇𝟐𝟐 𝒚𝒚
𝑖𝑖1 1.5 11 1
𝑖𝑖2 3.0 7 0
𝑖𝑖3 1.0 12 1
𝑖𝑖4 2.6 9 1
𝑖𝑖5 4.0 5 0

12
10 Minutes
Task 1.4 – Evaluation Basics
Assume you created a new machine learning algorithm. Discuss how you can evaluate
the performance of an algorithm on a dataset:

a) Given the predictions of a model and the true values, describe an approach to
evaluate the quality of the model’s predictions. (Assume we are talking about
regression)
b) You are given a dataset with 100 instances. You have trained (i.e., learned /
optimized) a model with an algorithm on the 100 instances. Now you want to test
the model on the same 100 instances. Is this a good idea?
c) Describe a valid approach to use 100 instances to train and evaluate (test) an
algorithm.

19
Exercise 1 should
already be uploaded.
Thank You and Goodbye!

Contact
• https://isg.beel.org/people/

• https://isg.beel.org/contact/

24

You might also like