Archetypal Analysis and DEA Model, Their Application on Financial Data and Visualization with PHATE

Area of Science:

• Economics
• Data Science
• Financial Analysis

Background:

• Macroeconomic analysis requires ranking and segmenting enterprises based on financial indicators.
• Company segmentation is crucial for developing effective investment strategies and evaluating financial risk.
• Visualizing complex financial data in a low-dimensional space aids in understanding enterprise relationships.

Purpose of the Study:

• To distinguish and group similar companies using financial indicators.
• To visualize these company clusters in a low-dimensional space for better insight.
• To compare the effectiveness of Data Envelopment Analysis (DEA) BCC model and archetypal analysis for enterprise segmentation.

Main Methods:

• Applied the Data Envelopment Analysis (DEA) BCC model to a dataset of companies listed on the Warsaw Stock Exchange.
• Utilized archetypal analysis to identify underlying structures and group similar companies.
• Employed the PHATE algorithm for dimensionality reduction and visualization of the analysis results.

Main Results:

• Both DEA BCC model and archetypal analysis yielded consistent results in segmenting companies.
• The PHATE algorithm successfully visualized the identified company clusters in a low-dimensional space.
• The study provided a clearer understanding of the data structure and inter-company similarities.

Conclusions:

• DEA BCC and archetypal analysis are effective methods for segmenting companies based on financial indicators.
• PHATE algorithm offers a valuable tool for visualizing complex financial data and analysis outcomes.
• The integrated approach enhances the ability to perform macroeconomic analysis, inform investment strategies, and assess risk.