Goto

Collaborating Authors

Improving Label Ranking Ensembles using Boosting Techniques

arXiv.org Machine Learning

Label ranking is a prediction task which deals with learning a mapping between an instance and a ranking (i.e., order) of labels from a finite set, representing their relevance to the instance. Boosting is a well-known and reliable ensemble technique that was shown to often outperform other learning algorithms. While boosting algorithms were developed for a multitude of machine learning tasks, label ranking tasks were overlooked. In this paper, we propose a boosting algorithm which was specifically designed for label ranking tasks. Extensive evaluation of the proposed algorithm on 24 semi-synthetic and real-world label ranking datasets shows that it significantly outperforms existing state-of-the-art label ranking algorithms.


Random Forest for Label Ranking

arXiv.org Machine Learning

Label ranking aims to learn a mapping from instances to rankings over a finite number of predefined labels. Random forest is a powerful and one of the most successfully general-purpose machine learning algorithms of modern times. In the literature, there seems no research has yet been done in applying random forest to label ranking. In this paper, We present a powerful random forest label ranking method which uses random decision trees to retrieve nearest neighbors that are not only similar in the feature space but also in the ranking space. We have developed a novel two-step rank aggregation strategy to effectively aggregate neighboring rankings discovered by the random forest into a final predicted ranking. Compared with existing methods, the new random forest method has many advantages including its intrinsically scalable tree data structure, highly parallel-able computational architecture and much superior performances. We present extensive experimental results to demonstrate that our new method achieves the best predictive accuracy performances compared with state-of-the-art methods for datasets with complete ranking and datasets with only partial ranking information.


Empirical Evaluation of Ranking Trees on Some Metalearning Problems

AAAI Conferences

The problem of learning rankings is receiving increased attention from several research communities. In this paper we empirically evaluate an adaptation of the algorithm of learning decision trees for rankings. Our experiments are carried out on some metalearning problems, which consist of relating characteristics of learning problems to the relative performance of learning algorithms. We obtain positive results which, somewhat surprisingly, indicate that the method predicts more accurately the top ranks.


Mining Rank Data

arXiv.org Machine Learning

The problem of frequent pattern mining has been studied quite extensively for various types of data, including sets, sequences, and graphs. Somewhat surprisingly, another important type of data, namely rank data, has received very little attention in data mining so far. In this paper, we therefore addresses the problem of mining rank data, that is, data in the form of rankings (total orders) of an underlying set of items. More specifically, two types of patterns are considered, namely frequent rankings and dependencies between such rankings in the form of association rules. Algorithms for mining frequent rankings and frequent closed rankings are proposed and tested experimentally, using both synthetic and real data.


Label Ranking with Abstention: Predicting Partial Orders by Thresholding Probability Distributions (Extended Abstract)

arXiv.org Artificial Intelligence

We consider an extension of the setting of label ranking, in which the learner is allowed to make predictions in the form of partial instead of total orders. Predictions of that kind are interpreted as a partial abstention: If the learner is not sufficiently certain regarding the relative order of two alternatives, it may abstain from this decision and instead declare these alternatives as being incomparable. We propose a new method for learning to predict partial orders that improves on an existing approach, both theoretically and empirically. Our method is based on the idea of thresholding the probabilities of pairwise preferences between labels as induced by a predicted (parameterized) probability distribution on the set of all rankings.