Category Archives: Machine Learning

Successful Presentations @useR!2017 Brussels

This slideshow requires JavaScript.

Last week the useR! 2017 conference took place in Brussels, Belgium. The annual useR! conference is the main meeting of the international R user and developer community. Over 1000 participants came to listen to a broad spectrum of talks ranging from technical and R-related computing issues to general statistical topics of current interest .

Here you find some impressions from the conference: Pictures

The SPOTSeven team gave two presentations:

Here you can find the slides:

The talks enjoyed great participation and resulted in lots of interesting discussions afterwards.

Maschinelles Lernen optimiert @TH_Koeln

v.l. Prof. Dr. Michael Bongards, Prof. Dr. Rainer Scheuring, Prof. Dr. Thomas Bartz-Beielstein, Jesiele N. de Lima (Eaton), Alexis Marino Sarda Espinosa, Andreas von der Beeck, Georg Reidt (Eaton), Prof. Dr. Christian Averkamp, Pia Müller (Eaton) ​(Bild: Manfred Stern / TH Köln)

Mit dem „Eaton-Award“, einem Förderpreis mit der Gesamtsumme von 3000 Euro, zeichnet der internationale Elektronik-Konzern Eaton jedes Jahr Bestleistungen der Studiengänge Elektronik und Automatisierungstechnik am Cam­pus Gummersbach der TH Köln aus.
Der zweite Preis mit 1.000 Euro Preisgeld ging in diesem Jahr an einen Mexikaner. Alexis Marino Sarda Espinosa schrieb seine Masterarbeit im englischsprachigen Studiengang “Automation & IT” mit dem Titel „A Machine Learning Approach for the Knowledge Extraction and Exploitation of Fleet Data“. Die Arbeit entstand für den internationalen Konzern ABB AG und wurde gleichermaßen vom Betreuer aus der Industrie, Dr. Subanatarajan Subbiah, und von Prof. Dr. Thomas Bartz-Beielstein vom Campus Gummersbach als “ausgezeichnet” beurteilt. In der Arbeit werden Verfahren des maschinellen Lernens analysiert und für den industriellen Einsatz angepasst. Continue reading

New Article: Conditional Inference Trees for the Knowledge Extraction from Motor Health Condition Data #ComputationalIntelligence #MachineLearning

The article “Conditional Inference Trees for the Knowledge Extraction from Motor Health Condition Data” (Alexis Sardá-Espinosaa, Subanatarajan Subbiah, Thomas Bartz-Beielstein), which will be published in the journal “Engineering Applications of Artificial Intelligence“, can be freely downloaded until May 20, 2017 from
Anyone who clicks on the link until May 20, 2017, will be taken to the final version of your article on ScienceDirect for free. No sign up or registration is needed – just click and read!

Abstract: Computational tools for the analysis of data gathered by monitoring systems are necessary because the amount of data steadily increases. Machine learning algorithms can be used in both regression and classification problems, providing useful insights while avoiding the bias and proneness to errors of humans. In this paper, a specific kind of decision tree algorithm, called conditional inference tree, is used to extract relevant knowledge from data that pertains to electrical motors. The model is chosen due to its flexibility, strong statistical foundation, as well as great capabilities to generalize and cope with problems in the data. The obtained knowledge is organized in a structured way and then analyzed in the context of health condition monitoring. The final results illustrate how the approach can be used to gain insight into the system and present the results in an understandable, user-friendly manner.

Keywords: Decision tree; Conditional inference tree; Health condition monitoring; Machine learning; Knowledge extraction

Authors: Alexis Sardá-Espinosa (ABB AG German Research Center, Technische Hochschule Köln),  Subanatarajan Subbiah (ABB AG German Research Center), Thomas Bartz-Beielstein (Technische Hochschule Köln)

Here is the DOI: 10.1016/j.engappai.2017.03.008



#Surrogate #Metamodels, and #MachineLearning article online: Model-based Methods for Continuous and Discrete Global #Optimization

The article “Model-based Methods for Continuous and Discrete Global Optimization” by T. Bartz-Beielstein and M. Zaefferer is available online:
A preprint can be downloaded from “Cologne Open Science”: urn:nbn:de:hbz:832-cos4-4356
The use of surrogate models is a standard method for dealing with complex real-world optimization problems. The first surrogate models were applied to continuous optimization problems. In recent years, surrogate models gained importance for discrete optimization problems. This article takes this development into consideration. The first part presents a survey of model-based methods, focusing on continuous optimization. It introduces a taxonomy, which is useful as a guideline for selecting adequate model-based optimization tools. The second part examines discrete optimization problems. Here, six strategies for dealing with discrete data structures are introduced. A new approach for combining surrogate information via stacking is proposed in the third part. The implementation of this approach will be available in the open source R package SPOT2. The article concludes with a discussion of recent developments and challenges in continuous and discrete application domains.