Scribd
Upload
12 views

Lab 4

Uploaded by

Bekhedda Asma
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
12 views

Lab 4

Uploaded by

Bekhedda Asma
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Data Mining

Lab sheet N°4: Dimensionality Reduction


ENSIA 2024-2025

Objectives
● Introduction to Dimensionality Reduction and its importance in reducing the number of variables
● Introduction to the Python library Scikit-learn for data analysis and modeling
● Use common dimensionality reduction techniques in Scikit-learn, including PCA
● Visualize the reduced data to gain insights into its characteristics
● Understand how to evaluate the effectiveness of dimensionality reduction

Required tools
Programming language: Python 3

Platforms: Anaconda, Jupyter Notebook, JupyterLab, Google Colab (cloud-based environment)

Python libraries:

● Scikit-learn: A data analysis and modeling library, including Data mining and Machine learning
algorithms for various tasks: classification, regression, clustering, dimensionality reduction, …

Resources
Scikit-learn:

● Reference card: Python For Data Science Cheat Sheet Scikit-Learn


● Official documentation: https://scikit-learn.org/stable/
● User guide: User guide: contents — scikit-learn 1.3.1 documentation
● Examples: Examples — scikit-learn 1.3.1 documentation
● Dataset loading utilities: 7. Dataset loading utilities — scikit-learn 1.3.1 documentation

PCA:

● Unsupervised reduction: https://scikit-learn.org/stable/modules/unsupervised_reduction.html


● Official documentation: sklearn.decomposition.PCA — scikit-learn 1.3.1 documentation
● Examples: PCA — scikit-learn 1.6.dev0 documentation

Notebooks
● Exercise: PCA from scratch.ipynb

● Guided Notebook: Principal_Component_Analysis_(Guided).ipynb

● Non-guided Notebook: Principal_Component_Analysis_(Non_Guided).ipynb


Part 1: Dimensionality reduction with Scikit-learn and PCA (150 minutes)
● Activate DM_ENV and install the required Python libraries: Scikit-learn
● Alternatively, students who find problems setting up the environment can use Google Colab
● Execute and understand the guided Jupyter Notebook file, in local or on Google Colab
● Fill in the gaps and write the missing code in the non-guided Jupyter Notebook file
● Take a look at the provided resources (documentation, tutorial, reference card) for more info

Part 2: Exercises on the Chapter Data - Part 2 (30 minutes)


1. How can a box plot be used to explore a data set containing four attributes: age, weight,
height, and income?

2. Explain how a box plot can indicate whether the values of an attribute are symmetrically
distributed.

3. What are the advantages of using a box plot over a histogram when exploring data
distributions, and in what scenarios would a histogram provide more insight?

4. How does the choice of the number and location of bins affect the appearance of a histogram,
and what methods can be used to reduce this dependency?

5. How can histograms, box plots, and density plots reveal information about skewness?
Additionally, what strategies can be employed to handle skewed data?

6. When should you use equal-width bins versus equal-frequency bins? What considerations
should be taken into account when making this choice?

7. How do you interpret the results of a quantile plot? What do straight lines and deviations from
the line indicate about the distribution of the data?

8. When might stratified sampling be more beneficial than simple random sampling, and how
does it influence the representation of subgroups within a data set?

You might also like