Instance%2DBased Ensemble Pruning via Multi%2DLabel Classification.

Obtaining Bipartitions from Score Vectors for Multi%2DLabel Classification.

Mining Multi%2Dlabel Data.

An ensemble uncertainty aware measure for directed hill climbing ensemble pruning.

Artificial Intelligence Applications and Innovations III

Transferring experience in reinforcement learning through task decomposition.
Transfer learning refers to the process of conveying experience from a simple task to another more complex (and related) task in order to reduce the amount of time that is required to learn the latter task. Typically, in a transfer learning procedure the agent learns a behavior in a source task, and it uses the gained knowledge in order to speed up the learning process in a target task. Reinforcement Learning algorithms are time expensive when they learn from scratch, especially in complex domains, and transfer learning comprises a suitable solution to speed up the training process. In this work we propose a method that decomposes the target task in several instances of the source task and uses them to extract an advised action for the target task. We evaluate the efficacy of the proposed approach in the robotic soccer Keepaway domain. The results demonstrate that the proposed method helps to reduce the training time of the target task.
On the Combination of Textual and Semantic Descriptions for Automated Semantic Web Service Classification.

Clustering Based Multi%2DLabel Classification for Image Annotation and Retrieval.

An Ensemble Pruning Primer.

An adaptive personalized news dissemination system.
Abstract. With the explosive growth of the Word Wide Web, information overload became a crucial concern. In a data-rich information-poor environment like the Web, the discrimination of useful or desirable information out of tons of mostly worthless data became a tedious task. The role of Machine Learning in tackling this problem is thoroughly discussed in the literature, but few systems are available for public use. In this work, we bridge theory to practice, by implementing a web-based news reader enhanced with a specifically designed machine learning framework for dynamic content personalization. This way, we get the chance to examine applicability and implementation issues and discuss the effectiveness of machine learning methods for the classification of real-world text streams. The main features of our system named PersoNews are: a) the aggregation of many different news sources that offer an RSS version of their content, b) incremental filtering, offering dynamic personalization of the content not only per user but also per each feed a user is subscribed to, and c) the ability for every user to watch a more abstracted topic of interest by filtering through a taxonomy of topics. PersoNews is freely available for public use on the WWW
Pruning an ensemble of classifiers via reinforcement learning.

Focused Ensemble Selection A Diversity%2DBased Method for Greedy Ensemble Selection.

An Ensemble of Classifiers for coping with Recurring Contexts in Data Streams.

An Empirical Study of Lazy Multilabel Classification Algorithms.

Multi%2DLabel Classification of Music into Emotions.

An Empirical Study of Lazy Multilabel Classification Algorithms.

Regression via Classification applied on software defect estimation.

Greedy regression ensemble selection Theory and an application to water quality prediction.

An empirical study on sea water quality prediction.

Random k %2DLabelsets An Ensemble Method for Multilabel Classification.
Abstract. This paper proposes an ensemble method for multilabel classification. The RAndom k-labELsets (RAKEL) algorithm constructs each member of the ensemble by considering a small random subset of labels and learning a single-label classifier for the prediction of each element in the powerset of this subset. In this way, the proposed algorithm aims to take into account label correlations using single-label classifiers that are applied on subtasks with manageable number of labels and adequate number of examples per label. Experimental results on common multilabel domains involving protein, document and scene classification show that better performance can be achieved compared to popular multilabel classification approaches. 1
An interoperable and scalable Web%2Dbased system for classifier sharing and fusion.

Multi%2DLabel Classification An Overview.
Nowadays, multi-label classification methods are increasingly required by modern applications, such as protein function classification, music categorization and semantic scene classification. This paper introduces the task of multi-label classification, organizes the sparse related literature into a structured presentation and performs comparative experimental results of certain multi-label classification methods. It also contributes the definition of concepts for the quantification of the multi-label nature of a data set.
Software Defect Prediction Using Regression via Classification.

PersoNews A Personalized News Reader Enhanced by Machine Learning and Semantic Filtering.
Abstract. In this paper, we present a web-based, machine-learning enhanced news reader (PersoNews). The main advantages of PersoNews are the aggregation of many different news sources, machine learning filtering offering personalization not only per user but also for every feed a user is subscribed to, and finally the ability for every user to watch a more abstracted topic of interest by employing a simple form of semantic filtering through a taxonomy of topics. 1.
Ensemble Pruning Using Reinforcement Learning.

On the Utility of Incremental Feature Selection for the Classification of Textual Data Streams.
Abstract. In this paper we argue that incrementally updating the features that a text classification algorithm considers is very important for real-world textual data streams, because in most applications the distribution of data and the description of the classification concept changes over time. We propose the coupling of an incremental feature ranking method and an incremental learning algorithm that can consider different subsets of the feature vector during prediction (what we call a feature based classifier), in order to deal with the above problem. Experimental results with a longitudinal database of real spam and legitimate emails shows that our approach can adapt to the changing nature of streaming data and works much better than classical incremental learning algorithms. 1
Protein Classification with Multiple Algorithms.
Abstract. Nowadays, the number of protein sequences being stored in central protein databases from labs all over the world is constantly increasing. From these proteins only a fraction has been experimentally analyzed in order to detect their structure and hence their function in the corresponding organism. The reason is that experimental determination of structure is labor-intensive and quite time-consuming. Therefore there is the need for automated tools that can classify new proteins to structural families. This paper presents a comparative evaluation of several algorithms that learn such classification models from data concerning patterns of proteins with known structure. In addition, several approaches that combine multiple learning algorithms to increase the accuracy of predictions are evaluated. The results of the experiments provide insights that can help biologists and computer scientists design high-performance protein classification systems of high quality. 1
HAPRC an automatically configurable planning system.

Selective fusion of heterogeneous classifiers.
There are two main paradigms in combining different classification algorithms: Classifier Selection and Classifier Fusion. The first one selects a single model for classifying a new instance, while the lat-ter combines the decisions of all models. The work presented in this paper stands in between these two paradigms aiming tackle the dis-advantages and benefit from the advantages of both. In particular, this paper proposes the use of statistical procedures for the selection of the best subgroup among different classification algorithms and the subsequent fusion of the decision of the models in this subgroup with simple methods like Weighted Voting. Extensive experimental results show that the proposed approach, Selective Fusion, improves over sim-ple selection and fusion methods, leading to performance comparable with the state-of-the-art heterogeneous classifier combination method of Stacking, without the additional computational cost and learning problems of meta-training. 1
Selective fusion of heterogeneous classifiers.
There are two main paradigms in combining different classification algorithms: Classifier Selection and Classifier Fusion. The first one selects a single model for classifying a new instance, while the lat-ter combines the decisions of all models. The work presented in this paper stands in between these two paradigms aiming tackle the dis-advantages and benefit from the advantages of both. In particular, this paper proposes the use of statistical procedures for the selection of the best subgroup among different classification algorithms and the subsequent fusion of the decision of the models in this subgroup with simple methods like Weighted Voting. Extensive experimental results show that the proposed approach, Selective Fusion, improves over sim-ple selection and fusion methods, leading to performance comparable with the state-of-the-art heterogeneous classifier combination method of Stacking, without the additional computational cost and learning problems of meta-training. 1
Effective Voting of Heterogeneous Classifiers.
This paper deals with the combination of classification models  that have been derived from running di#erent (heterogeneous) learning  algorithms on the same data set. We focus on the Classifier Evaluation  and Selection (ES) method, that evaluates each of the models (typically  using 10-fold cross-validation) and selects the best one. We examine  the performance of this method in comparison with the Oracle selecting  the best classifier for the test set and show that 10-fold cross-validation  has problems in detecting the best classifier. We then extend ES by applying  a statistical test to the 10-fold accuracies of the models and combining  through voting the most significant ones. Experimental results  show that the proposed method, E#ective Voting, performs comparably  with the state-of-the-art method of Stacking with Multi-Response Model  Trees without the additional computational cost of meta-training.
Using the k%2DNearest Problems for Adaptive Multicriteria Planning.
Abstract. This paper concerns the design and development of an adaptive planner that is able to adjust its parameters to the characteristics of a given problem and to the priorities set by the user concerning plan length and planning time. This is accomplished through the implementation of the k nearest neighbor machine learning algorithm on top of a highly adjustable planner, called HAP. Learning data are produced by running HAP offline on several problems from multiple domains using all value combinations of its parameters. When the adaptive planner HAPNN is faced with a new problem, it locates the k nearest problems, using a set of measurable problem characteristics, retrieves the performance data for all parameter configurations on these problems and performs a multicriteria combination, with user-specified weights for plan length and planning time. Based on this combination, the configuration with the best performance is then used in order to solve the new problem. Comparative experiments with the statistically best static configurations of the planner show that HAPNN manages to adapt successfully to unseen problems, leading to an increased planning performance. 1
Clustering classifiers for knowledge discovery from physically distributed databases.
Most distributed classification approaches view data distribution as a technical issue and combine local models aiming at a single global model. This however, is unsuitable for inherently distributed databases, which are often described by more than one classification models that might di#er conceptually. In this paper we present an approach for clustering distributed classifiers in order to discover groups of similar classifiers and thus similar databases with respect to a specific classification task. We also show that clustering distributed classifiers as a pre-processing step for classifier combination enhances the achieved predictive performance of the ensemble.
Learning Rules for Adaptive Planning.

Learning Rules for Adaptive Planning.

Effective Stacking of Distributed Classifiers.
One of the most promising lines of research towards discovering global predictive models from physically distributed data sets is local learning and model integration. Local learning avoids moving raw data around the distributed nodes and minimizes communication, coordination and synchronization cost. However, the integration of local models is not a straightforward process. Majority Voting is a simple solution that works well in some domains, but it does not always offer the best predictive performance. Stacking on the other hand, offers flexibility in modelling, but brings along the problem of how to train on sufficient and at the same time independent data without the cost of moving raw data around the distributed nodes. In addition, the scalability of Stacking with respect to the number of distributed nodes is another important issue that has not yet been substantially investigated. This paper presents a framework for constructing a global predictive model from local classifiers that does not require moving raw data around, achieves high predictive accuracy and scales up efficiently with respect to large numbers of distributed data sets.
