Cluster analysis of Wisconsin Breast Cancer dataset using self-organizing maps

Area of Science:

• Data Science
• Machine Learning
• Bioinformatics

Background:

• Multidimensional data analysis is crucial for extracting insights from complex datasets.
• Cluster analysis is a key technique for identifying patterns and groupings in data.
• The Wisconsin Breast Cancer dataset is a widely used benchmark for classification and clustering tasks.

Purpose of the Study:

• To apply cluster analysis and Kohonen self-organizing maps to the Wisconsin Breast Cancer dataset.
• To evaluate the effectiveness of self-organizing maps for multidimensional data assessment.
• To critically assess the common usage of the Wisconsin Breast Cancer dataset for benchmarking.

Main Methods:

• The study explains the concept of cluster analysis and compares different methods.
• It describes the Kohonen model of self-organizing maps and its application to cluster analysis.
• The Wisconsin Breast Cancer dataset was analyzed using these methodologies.

Main Results:

• Cluster analysis using self-organizing maps provided a useful complement for assessing the multidimensional Wisconsin Breast Cancer data.
• Visual and textual descriptions of the findings were generated.
• The analysis highlighted patterns within the dataset.

Conclusions:

• Self-organizing maps offer a valuable approach for exploring and understanding complex datasets.
• The Wisconsin Breast Cancer dataset has been overutilized for automated decision benchmarking without sufficient in-depth data analysis.
• Further research should focus on comprehensive data analysis rather than solely benchmarking.