Skip to content

Repository to work on clustering exercises using machine learning

Notifications You must be signed in to change notification settings

jamesallen0351/clustering-exercises

Repository files navigation

clustering-exercises

  • This repository is for my clustering exercises and notes.

Clustering

Learning Goals

  1. Be able to discuss and discover use cases for clustering across multiple industries.

  2. Be able to recognize of common clustering algorithms.

  3. General understanding of how the k-means clustering algorithm works.

  4. Ability to implement k-means clustering in python.

  5. Handle outliers using IQR.

  6. More practice scaling data.

  7. Strategies for missing values.

  8. Plotting clusters.

  9. Ability to make use of clusters discovered later down the data science pipeline.

Data Wrangling: Acquisition & Preparation

Learning Goals

  1. Acquire a sample of data from SQL.

  2. Identify null values, which nulls are 'deal-breakers', i.e. rows removed, which nulls should be represented by 0, and which

  3. should be replaced by a value from other methods, such as mean.

  4. Identify outliers and decide what to do with them, if anything (remove, keep as-is, replace).

  5. Data Structure: Aggregate as needed so that every row is an observation and each column is a variable (1 variable and not a measure).

Exploratory Analysis

"Exploratory Data Analysis refers to the critical process of performing initial investigations on data so as to discover patterns,to spot anomalies,to test hypothesis and to check assumptions with the help of summary statistics and graphical representations." - Prasad Patil, Novice Data Science Storyteller, Mar 23

Goals of exploration

  1. Can we see patterns, find signals in the data?

  2. What features are driving the outcome?

  3. Are there other features we can construct that have stronger relationships?

  4. Use Visualization and statistical testing to help answer these questions.

  5. We want to walk away from exploration with with modeling strategies (feature selection, algorithm selection, evaluation methods, for example).

K-Means

At A Glance

  • Unsupervised machine learning algorithm
  • Groups a dataset into a user-specified number (k) of clusters
  • k must be specified ahead of time
  • clusters won't have the same number of points

Using Clusters

Explore your Clusters

Turn your Clusters into Labels

Model Each Cluster Separately

Turn your Clusters into Features

About

Repository to work on clustering exercises using machine learning

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published