Integrated Dual Analysis of Quantitative and Qualitative High-Dimensional Data.

Area of Science:

• Data Science
• Medical Informatics
• Bioinformatics

Background:

• High-dimensional data exploration is challenging.
• Existing frameworks often focus on quantitative data.
• Integrating qualitative data is crucial for comprehensive analysis.

Purpose of the Study:

• Extend the Dual Analysis framework to jointly analyze quantitative and qualitative data.
• Develop methods for visualizing and exploring mixed-type, high-dimensional datasets.
• Demonstrate the framework's utility in generating novel hypotheses from complex data.

Main Methods:

• Adapted measures of variation for qualitative data (nominal, ordinal) to be compatible with quantitative data.
• Integrated these measures into the Dual Analysis framework.
• Enabled joint visualization and interactive exploration of mixed data types.
• Applied the extended framework to a medical case study on Cerebral Small Vessel Disease (CSVD).

Main Results:

• The extended Dual Analysis framework successfully visualizes and analyzes mixed quantitative and qualitative dimensions together.
• Common measures of variation facilitate joint data treatment during interactive exploration.
• The approach allows for rapid hypothesis generation from high-dimensional mixed data.
• A case study on CSVD data yielded new insights, supporting hypothesis formulation.

Conclusions:

• The joint treatment of quantitative and qualitative data within the Dual Analysis framework enhances exploratory analysis.
• This integrated approach accelerates the discovery of new insights and hypothesis generation.
• The framework shows significant potential for applications in medical research and other fields dealing with complex, mixed-type data.