Skip to content

Utilized machine learning algorithms (K-Means, Birch, Gaussian Mixture Model, Hierarchical Clustering, Mean-Shift Clustering) to identify hotspots. Validation with Silhouette Score and Heatmap. Hierarchical Clustering scored 0.943, demonstrating exceptional performance.

Notifications You must be signed in to change notification settings

swarnima000/Hotspot-Detection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 

Repository files navigation

Hotspot-Detection

Description

The Hotspot Detector project is a machine learning-based initiative aimed at identifying hotspots within a defined geographic region. Leveraging a variety of algorithms, including K-Means, Birch, Gaussian Mixture Model, Hierarchical Clustering, and Mean-Shift Clustering, the project employs sophisticated clustering techniques to pinpoint areas of interest. Validation of the algorithmic results is conducted using robust metrics such as the Silhouette Score and Heatmap.

Uses

This model can be utilized as per specific requirements, including:

  • Crime Hotspots: Identifying areas with high crime rates to optimize the deployment of security services.
  • Taxi Hotspots: Identifying locations with high demand for taxis but limited supply, enabling the strategic placement of taxi drivers.

Accuracy

The Hierarchical Clustering algorithm exhibited outstanding performance, achieving a silhouette score of 0.943. This indicates the model's exceptional ability to accurately identify and delineate hotspots within the specified area.

Used Dataset

https://www.kaggle.com/datasets/gpreda/coronavirus-2019ncov

About

Utilized machine learning algorithms (K-Means, Birch, Gaussian Mixture Model, Hierarchical Clustering, Mean-Shift Clustering) to identify hotspots. Validation with Silhouette Score and Heatmap. Hierarchical Clustering scored 0.943, demonstrating exceptional performance.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published