Papers by Frederik Janssen
The primary goal of the research reported in this paper is to identify what criteria are responsi... more The primary goal of the research reported in this paper is to identify what criteria are responsible for the good performance of a heuristic rule evaluation function in a greedy topdown covering algorithm. We first argue that search heuristics for inductive rule learning algorithms typically trade off consistency and coverage, and we investigate this trade-off by determining optimal parameter settings for five different parametrized heuristics. In order to avoid biasing our study by known functional families, we also investigate the potential of using meta-learning for obtaining alternative rule learning heuristics. The key results of this experimental study are not only practical default values for commonly used heuristics and a broad comparative evaluation of known and novel rule learning heuristics, but we also gain theoretical insights into factors that are responsible for a good performance. For example, we observe that consistency should be weighed more heavily than coverage, presumably because a lack of coverage can later be corrected by learning additional rules.
Nachweis von Sojaeiwei� in erhitzten Fleischerzeugnissen durch elektrophoretischen Transfer und T�pfeltest
Eur Food Res Technol, 1986
Zeitschrift Fur Le0bensmittel Untersuchung Und Forschung, Jan 5, 1986
Zusammenfassung. Sojaprotein (Isolat, Konzentrat, Texturat) und Fleischerzeugnisse mit Sojazusatz... more Zusammenfassung. Sojaprotein (Isolat, Konzentrat, Texturat) und Fleischerzeugnisse mit Sojazusatz wurden einer SDS-Elektrophorese unterworfen. Nachher wurde auf Nitrocellulose elektrophoretisch transferiert und selektiv mit einer Immunoperoxydase-Methode angef/irbt, und zwar in folgender Reihenfolge: prim/ires Kaninchen-Antisojaserum, Antikaninchen-IG-Serum vonder Ziege und einem Peroxydase-Antiperoxydase-Komplex (Kaninchen-Allotyp). Nach Inkubation mit Peroxidasesubstrat konnte gezeigt wetden, dab alle (reduzierten) Sojafraktionen @% ~'-,/% Conglycinin wie auch die saure und basische Glycininfraktion) angef/irbt wurden. Das auf diese Weise erhaltene Muster hat ffir die Anwesenheit fiir Sojaprotein in Fleischerzeugnissen also eine hohe Aussagekraft im Vergleich mit einer unspezifischen F/irbung, wie zum Beispiel mit Coomassie BB R250. Die Nachweisgrenze liegt bei etwa 0,02% Sojaisolat.
European Journal of Gastroenterology Hepatology, Jun 1, 2001
Meta-Learning a Rule Learning Heuristic
Methods to analyse epicardial mapping signals for theelectrophysiological detection of viability in chronic ischemicmyocardium
ABSTRACT
Non-tidal oceanic mass shifts and their impact on terrestrial gravity observations
In the last years the application of high-quality terrestrial gravity observations to studies of ... more In the last years the application of high-quality terrestrial gravity observations to studies of mass transport has become more and more a focus of interest. This goes hand in hand with efforts to enhance the signal-to-noise ratios in the time series by more comprehensive and sophisticated reductions for geodynamic studies. In addition a precondition for a consistent combination of terrestrial
Evaluation metrics for rule learning typically, in one way or another, trade off consistency and ... more Evaluation metrics for rule learning typically, in one way or another, trade off consistency and coverage. In this work, we investigate this trade- off for three different families of rule learn- ing heuristics, all of them featuring a parame- ter that implements this trade-off in different guises. These heuristics are the m-estimate, the F -measure, and the Klösgen measures. The main goals of this work are to extend our understand- ing of these heuristics by visualizing their behav- ior via isometrics in coverage space, and to deter- mine optimal parameter settings for them. Inter- estingly, even though the heuristics use quite dif- ferent ways for implementing this trade-off, their optimal settings realize quite similar evaluation functions. Our empirical results on a large num- ber of datasets demonstrate that, even though we do not use any form of pruning, the quality of the rules learned with these settings outperforms standard rule learning heuristics and approaches the perfor...
Most commonly used inductive rule learning algorithms employ a hill-climbing search, whereas loca... more Most commonly used inductive rule learning algorithms employ a hill-climbing search, whereas local pattern discovery algorithms employ exhaustive search. In this paper, we evaluate the spectrum of different search strategies to see whether separate-and-conquer rule learning algorithms are able to gain performance in terms of predictive accuracy or theory size by using more powerful search strategies like beam search or exhaustive search. Unlike previous results that demonstrated that rule learning algorithms suffer from over-searching, our work pays particular attention to the interaction between the search heuristic and the search strategy. Our results show that exhaustive search has primarily the effect of finding longer, but nevertheless more general rules than hill-climbing search. Thus, in cases where hillclimbing finds too specific rules, exhaustive search may help, while in others it may lead to over-generalization. 1
Ontology matching approaches have mostly worked on the schema level so far. With the advent of Li... more Ontology matching approaches have mostly worked on the schema level so far. With the advent of Linked Open Data and the avail-ability of a massive amount of instance information, instance-based ap-proaches become possible. This position paper discusses approaches and challenges for using those instances as input for machine learning algo-rithms, with a focus on rule learning algorithms, as a means for ontology matching.
Intelligente Datenauswertung mit Linked Open Data
X.media.press, 2015
Accurate pollutant modeling and mapping: Applying machine learning to participatory sensing and urban topology data
2015 International Conference and Workshops on Networked Systems (NetSys), 2015
Separating Rule Refinement and Rule Selection Heuristics in Inductive Rule Learning
Lecture Notes in Computer Science, 2014
Proceedings of the 2009 SIAM International Conference on Data Mining, 2009
Lecture Notes in Computer Science, 2008
In this paper, we argue that search heuristics for inductive rule learning algorithms typically t... more In this paper, we argue that search heuristics for inductive rule learning algorithms typically trade off consistency and coverage, and we investigate this trade-off by determining optimal parameter settings for five different parametrized heuristics. This empirical comparison yields several interesting results. Of considerable practical importance are the default values that we establish for these heuristics, and for which we show that they outperform commonly used instantiations of these heuristics. We also gain some theoretical insights. For example, we note that it is important to relate the rule coverage to the class distribution, but that the true positive rate should be weighted more heavily than the false positive rate. We also find that the optimal parameter settings of these heuristics effectively implement quite similar preference criteria.
European journal of gastroenterology & hepatology, 2001
The prolamin working group coordinates research on laboratory gluten analysis in food and on clin... more The prolamin working group coordinates research on laboratory gluten analysis in food and on clinical evaluation of patient sensitivity to prolamins. As an observer organization to the Codex Alimentarius Commission, the group summarizes current data on analysis and effects of gluten in coeliac disease. All types of gliadin, the ethanol-soluble fraction of gluten, contain the coeliac-active factor. However, coeliac toxicity and immunogenicity (humoral and cellular) of various prolamins are not identical in coeliac patients. There are no conclusive data on the threshold of gluten sensitivity of coeliac patients. Information as to the long-term risk to coeliac patients exposed to small doses of gliadin is lacking. Therefore, every effort should be made to keep the diet of coeliac patients as gluten-free as possible. The prolamin group is currently evaluating a new enzyme-linked immunosorbent assay (ELISA) protocol for gluten analysis that could serve as a basis for further Codex regula...
Towards Multilabel Rule Learning
Semantic Abstraction for Generalization of Tweet Classification
Data-driven Detection of Congestion-affected Roads
Semantic Abstraction for Generalization of Tweet Classification: An Evaluation on Incident-Related Tweets
Uploads
Papers by Frederik Janssen