mv-streaming-algorithms

Algorithms developed for Kesavan et al., "A Visual Analytics Framework for Reviewing Streaming Performance Data", 2020.

Algorithms for Multivariate Streaming Data Analysis

About

• Algorithms for multivariate streaming data analysis. We originally developed these algorithms for : A Visual Analytics Framework for Reviewing Streaming Performance Data. Suraj P. Kesavan, Takanori Fujiwara, Jianping Kelvin Li, Caitlin Ross, Misbah Mubarak, Christopher D. Carothers, Robert B. Ross, and Kwan-Liu Ma. In Proceedings of IEEE Pacific Visualization Symposium (PacificVis), forthcoming

• Demonstration of a system using the algorithms: https://youtu.be/pxthZSJ1jqs.

  Also, you can find source code of the demonstrated system: https://github.com/VIDILabs/Streaming-ROSS-Project.

• Features

  • Online and progressive algorithms for multivariate streaming data analysis. These algorithms can provide results with required latency.

    1. Incremental PCA-based online change point detection with adaptive forgetting factor.
    2. Progressive PCA using incremental mechanism with consideration of visual consistency.
    3. Progressive k-means clustering with with consideration of visual consistency.
    4. Progressive causality analysis methods

  • For details and references for each algorithm, please see the corresponding subdirectories.

---

Requirements and How to Set Up

• Please refer to README.md placed in each corresponding subdirectory.

---

How to Cite

Please, cite:
Suraj P. Kesavan, Takanori Fujiwara, Jianping Kelvin Li, Caitlin Ross, Misbah Mubarak, Christopher D. Carothers, Robert B. Ross, and Kwan-Liu Ma. "A Visual Analytics Framework for Reviewing Streaming Performance Data." In Proceedings of IEEE Pacific Visualization Symposium (PacificVis), forthcoming

---

License

Because some method is developed with external libraries, each method has a different license. Please, refer to LICENSE file in each corresponding subdirectory.