The utility of clusters and a Hungarian clustering algorithm.

Area of Science:

• Computer Science
• Data Mining
• Machine Learning

Background:

• The k-means algorithm assigns cluster utility based on the sum of distances from the centroid.
• Determining the optimal number of clusters (k) in k-means can be challenging.
• Existing methods lack a natural way to evaluate and select clusters sequentially.

Purpose of the Study:

• To introduce an alternative method for assigning utility to clusters.
• To provide a framework for determining the optimal number of clusters (k) in a sequential manner.
• To enhance cluster analysis by offering a new evaluation metric.

Main Methods:

• Developing an alternative cluster utility assignment method.
• Utilizing optimizations over permutations.
• Employing cyclic groups and the Hungarian algorithm for cluster generation.

Main Results:

• A new, effective method for cluster evaluation has been proposed.
• The method facilitates sequential identification of optimal clusters.
• Demonstrated utility in determining an appropriate number of clusters (k).

Conclusions:

• The proposed cluster evaluation method offers a valuable alternative to traditional k-means utility.
• This approach provides a robust solution for determining the optimal number of clusters.
• The method enhances the flexibility and applicability of clustering algorithms.