Types of Clusters and Objective of Function - Clustering - Data Mining and Business Intelligence

TL;DR

This video discusses the different types of clusters (center-based, contagious, density-based, conceptual) and their characteristics, as well as the importance of objective functions in clustering algorithms.

Transcript

hello friends in this video we are going to see about the types of clusters as well as the object of a functions why we are performing basically the question and how we can work it so first we'll see the types of clusters so basically there are five types of clusters first is Center Bay second is contagious cluster density based luster for this con... Read More

Key Insights

• ⚾ There are five types of clusters: center-based, contagious, density-based, conceptual, and described by objective functions.
• 😥 Center-based clusters divide data based on centroids, while contagious clusters group points closer to each other than to points in other clusters.
• 😘 Density-based clusters have regions of high density separated by low-density regions, while conceptual clusters share specific characteristics.
• 🆘 Objective functions help evaluate the quality of clusters by minimizing or maximizing certain measures.
• 🌐 Global objective functions consider partitioning clustering algorithms, while local objective functions consider hierarchical clustering algorithms and density-based clustering algorithms.

Questions & Answers

Q: What is the difference between center-based and density-based clusters?

Center-based clusters divide data based on centroids, while density-based clusters identify regions of high density separated by regions of low density. Center-based clusters involve dividing data using specific centers, while density-based clusters focus on the density of the data.

Q: How are contagious clusters formed?

Contagious clusters are formed by grouping points that are closer to each other than to any points in other clusters. Each point is at least closer to one point in its cluster than to any point in other clusters.

Q: What are conceptual clusters?

Conceptual clusters involve points sharing a certain characteristic or property derived from the entire set of points. They are not center-based or density-based, making them harder to detect.

Q: What is the role of objective functions in clustering?

Objective functions help determine the quality of clusters by minimizing or maximizing specific measures. One common objective function is the minimization of the sum of square error, which calculates the distance between data points and the center of their cluster.

Summary & Key Takeaways

• There are five types of clusters: center-based, contagious, density-based, and conceptual clusters.

• Center-based clusters divide data based on centroids, with data closer to the center of a cluster than any other cluster.

• Contagious clusters have each point closer to one point in its cluster than any other cluster.

• Density-based clusters have regions of high density separated by regions of low density.