Convergent Analytics Convergent_Analytics
← Back to Papers
Journal Article University of Cincinnati CEAS

Computation of Persistent Homology on Streaming Data using Topological Data Summaries

Anindya Moitra, Nicholas O. Malott, Philip A. Wilsey
| Computational Intelligence
Persistent HomologyStreaming DataTopological Data Analysis

Summary

Methods for computing persistent homology in streaming environments through topological summaries that reduce computational cost while preserving meaningful topological structure.

Research Context

This publication reflects graduate research conducted at the University of Cincinnati College of Engineering and Applied Science. It is included here to document the technical foundation behind Convergent Analytics' work in industrial analytics, applied AI, high-performance computing, and topological data analysis.

This peer-reviewed journal publication originated from graduate research conducted at the University of Cincinnati College of Engineering and Applied Science.

The work addresses one of the major limitations of traditional persistent homology algorithms: the assumption that all data is available prior to computation. Many real-world systems generate data continuously, requiring new approaches that can process topological information incrementally.

The paper introduces methods for summarizing streaming datasets while preserving topological characteristics, enabling scalable analysis of evolving data sources and laying the foundation for future applications in monitoring, sensing, and industrial analytics environments.

Citation

Moitra, A., Malott, N.O., Wilsey, P.A. Computation of Persistent Homology on Streaming Data using Topological Data Summaries. Computational Intelligence. DOI: 10.1111/coin.12597