2012
Authors
Nassif, H; Santos Costa, V; Burnside, ES; Page, D;
Publication
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Abstract
A typical classification problem involves building a model to correctly segregate instances of two or more classes. Such a model exhibits differential prediction with respect to given data subsets when its performance is significantly different over these subsets. Driven by a mammography application, we aim at learning rules that predict breast cancer stage while maximizing differential prediction over age-stratified data. In this work, we present the first multi-relational differential prediction (aka uplift modeling) system, and propose three different approaches to learn differential predictive rules within the Inductive Logic Programming framework. We first test and validate our methods on synthetic data, then apply them on a mammography dataset for breast cancer stage differential prediction rule discovery. We mine a novel rule linking calcification to in situ breast cancer in older women. © 2012 Springer-Verlag.
2012
Authors
Fonseca, NA; Santos Costa, V; Camacho, R;
Publication
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Abstract
"Traditional" clustering, in broad sense, aims at organizing objects into groups (clusters) whose members are "similar" among them and are "dissimilar" to objects belonging to the other groups. In contrast, in conceptual clustering the underlying structure of the data together with the description language which is available to the learner is what drives cluster formation, thus providing intelligible descriptions of the clusters, facilitating their interpretation. We present a novel conceptual clustering system for multi-relational data, based on the popular k?-?medoids algorithm. Although clustering is, generally, not straightforward to evaluate, experimental results on several applications show promising results. Clusters generated without class information agree very well with the true class labels of cluster's members. Moreover, it was possible to obtain intelligible and meaningful descriptions of the clusters. © 2012 Springer-Verlag Berlin Heidelberg.
2012
Authors
Acar, UA; Costa, VS;
Publication
DAMP
Abstract
2012
Authors
Dovier, A; Costa, VS;
Publication
ICLP (Technical Communications)
Abstract
2012
Authors
Acar, U; Costa, VS;
Publication
Conference Record of the Annual ACM Symposium on Principles of Programming Languages
Abstract
2012
Authors
Ferreira, CA; Gama, J; Costa, VS;
Publication
COMPUTER AND INFORMATION SCIENCES II
Abstract
In this work we present XmuSer, a multi-relational framework suitable to explore temporal patterns available in multi-relational databases. xMuS er's main idea consists of exploiting frequent sequence mining, using an efficient and direct method to learn temporal patterns in the form of sequences. Grounded on a coding methodology and on the efficiency of sequence miners, we find the most interesting sequential patterns available and then map these findings into a new table, which encodes the multi-relational timed data using sequential patterns. In the last step of our framework, we use an ILP algorithm to learn a theory on the enlarged relational database that consists on the original multi-relational database and the new sequence relation. We evaluate our framework by addressing three classification problems.
The access to the final selection minute is only available to applicants.
Please check the confirmation e-mail of your application to obtain the access code.