Master Thesis on time series analysis expands functionality of BOSS

06 February 2018

BOSS is Solenix’s web-based real-time dashboard solution for intelligent monitoring. Based on the Elasticsearch technology stack, we have further enhanced the offered data analysis capabilities with the addition of a scripting engine, which allows the integration of data mining algorithms.

As part of this feature, Solenix has conducted several research studies and collaborations with universities to explore different possibilities to offer novel and robust data insights. The last conducted Master Thesis was that of TU Darmstadt student Elvir Sabic and focussed on “Finding dependencies between time series in satellite data”. The thesis, which ran from March 2017 to September 2017, applied an information theoretic measure called transfer entropy to quantify the relationship between different time series. The findings show that this method may be useful for anomaly and fault detection.

The results of this Master Thesis complement previous projects and studies on this and other related topics, such as outlier and anomaly detection and their classification. These results are constantly being integrated in our BOSS solution, making it more complete, robust and reliable and offering new insights and perspectives on the monitored data.

Image Credits: Master-Thesis von Elvir Sabic, Finden von Abhängigkeiten zwischen Zeitreihen in Satellitendaten, September 2017. TU Darmstadt.