KDD Cup 2001 GeneProtein Function Prediction Using the

KDD Cup 2001: Gene/Protein Function Prediction Using the Multirelational Learning Algorithm RELAGGS Mark-A. Krogel Otto-von-Guericke-Universität, Magdeburg, Germany School of Computer Science, Research Group Stefan Wrobel Mark-A. Krogel, Magdeburg University, Knowledge Discovery and Machine Learning Group 1

Preparation: A Multirelational Task o General: renormalize into multiple tables as a natural representation of the data genes_ relation gene class complex phenotype motif interactions_ relation interaction gene. N o Specific for KDD Cup tasks 2/3: consider only interactions with high correlations, assume transitivity, make symmetry explicit Mark-A. Krogel, Magdeburg University, Knowledge Discovery and Machine Learning Group 2

Algorithm: RELAGGS [Krogel/Wrobel: ILP 01] o Computes selected joins following user-defined foreign links o Performs automatic transformation of multiple tables into single table with the help of aggregate functions gene class complex phenotype RELAGGS gene_ for_ analysis motif interaction gene. N o Uses propositional learner such as C 4. 5 or SVMlight Mark-A. Krogel, Magdeburg University, Knowledge Discovery and Machine Learning Group 3

Summary o RELAGGS allows to work with natural multirelational form of data immediately o Easy specification of possible joins with foreign links o Maximal preservation of information through aggregation o Accuracies: 93, 6% on task 2: rank 1 69, 8% on task 3: rank 4 o http: //kd. cs. uni-magdeburg. de o M. -A. Krogel, S. Wrobel: Transformation-Based Learning Using Multirelational Aggregation. 11 th International Conference On Inductive Logic Programming, Strasbourg, France, Sept. 2001. Mark-A. Krogel, Magdeburg University, Knowledge Discovery and Machine Learning Group 4