Enriching Ordinal Factor Analysis with Algorithms for Visual Data Science

Ordinal data analysis is crucial for comprehending complex phenomena in various fields, including social sciences, biology, and more. However, current methods struggle to handle large datasets or non-linear structures, limiting their applicability. This article aims to address these limitations by developing a mature tool for ordinal factor analysis, enabling practitioners to analyze data with ease.
The research focuses on two key areas: improving the theoretical backbone and creating algorithms for efficient data analysis. The proposed methods will extract structure from data with minimal manual effort and handle incremental changes in the dataset. Additionally, a non-linear factor variant (e.g., tree-based factors) may be explored to further expand the tool’s capabilities.
To lay the groundwork for these advancements, the authors introduce an iterative greedy algorithm for computing linear factorizations in two or more dimensions. This algorithm builds on Ganter’s idea [70] to cover a maximum subset of the dataset with a small set of factors.
In summary, this article strives to make ordinal data analysis more accessible and practical by developing novel algorithms that can efficiently handle large datasets and non-linear structures. By doing so, researchers and practitioners will be able to gain deeper insights into complex phenomena in various fields, ultimately leading to new discoveries and advancements.