Please click here to read this text in the German language.

Please click here for the evaluation of the performance of the Databionic swarm in comparison to 26 common clustering algorithms using 15 datasets.

This site will provide you with the collection of knowledge in the field of data science. I have successfully applied data science methods (in the industry) for detecting anomalies in (working) IoT devices on the market (realm of big data using cloud computing) and bench tested devices during the production process, as well as for sales forecasting. I have worked on predictive maintenance of IoT devices in big data and made suggestions for improving the customer service by investigating a telephone routing system as well as for identifying unprofitable customers and regions based on scores. The scores were generated usin Gaussian mixture models.

My research interests are about methods derived from nature based on concepts like swarm intelligence, self-organization or emergence. The focus lies on unsupervised machine learning. Here approaches are presented on data visualization, dimensionality reduction, and cluster analysis with the goal to enable non-professionals in the field to apply data science to common problems by providing the relevant and application-focused information and functions in the language R.

I helped to extract new and interesting knowledge out datasets regarding municipalities, geographic variations, gene expressions, for counterfeit detection, out of pairs of alleles (codominant markers), chemical measurements, gross domestic products, using the gene ontology (GO) database, out of customer data, German peoples income, biomarkers, visual eye movements (fixations) from observers who viewed images, historical stock quotes, hyperspectral images, lipid marker serum concentrations in multiple sclerosis and many more.

My Book answers the question of how the artificial swarm can display connections in data using a 3D landscape. The concepts of self-organization, emergence, swarm intelligence and game theory are briefly outlined. These concepts serve to explain how the databionic swarm algorithm works. The artificial swarm is capable of automatic grouping of data and of showing relationships in a comprehensible way. The relevance of the method is demonstrated in the examples from cancer research, economics, hydrology and pain research. The databionic swarm is available free of charge as an open source project. Only a limited knowledge of computer science is required for an application.

For a broader spectrum of bionic methods in data science (e.g. emergent self-organizing maps, evolutionary algorithms, applications of self-regulating genes) I refer to databionics.