A fundamental problem when performing incremental learning is that the best set of a classification system's parameters can change with the evolution of the data. Consequently, unless the system self-adapts to such changes, it will become obsolete, even if the application environment seems to be static. To address this problem, we propose a dynamic optimization approach in this paper that performs incremental learning in an adaptive fashion by tracking, evolving, and combining optimum hypotheses overtime. The approach incorporates various theories, such as dynamic particle swarm optimization, incremental support vector machine classifiers, change detection, and dynamic ensemble selection based on classifiers' confidence levels. Experiments carried out on synthetic and real-world databases demonstrate that the proposed approach actually outperforms the classification methods often used in incremental learning scenarios. © 2011 Wiley Periodicals, Inc.

The incremental updating of classifiers implies that their internal parameter values can vary according to incoming data. As a result, in order to achieve high performance, incremental learner systems should not only consider the integration of knowledge from new data, but also maintain an optimum set of parameters. In this paper, we propose an approach for performing incremental learning in an adaptive fashion with an ensemble of support vector machines. The key idea is to track, evolve, and combine optimum hypotheses over time, based on dynamic optimization processes and ensemble selection. From experimental results, we demonstrate that the proposed strategy is promising, since it outperforms a single classifier variant of the proposed approach and other classification methods often used for incremental learning.

Support Vector Machines (SVM) are very powerful classifiers in theory but their efficiency in practice rely on an optimal selection of hyper-parameters. A naïve or ad hoc choice of values for the latter can lead to poor performance in terms of generalization error and high complexity of parameterized models obtained in terms of the number of support vectors identified. This hyper-parameter estimation with respect to the aforementioned performance measures is often called the model selection problem in the SVM research community. In this paper we propose a strategy to select optimal SVM models in a dynamic fashion in order to attend that when knowledge about the environment is updated with new observations and previously parameterized models need to be re-evaluated, and in some cases discarded in favour of revised models. This strategy combines the power of the swarm intelligence theory with the conventional grid-search method in order to progressively identify and sort out potential solutions using dynamically updated training datasets. Experimental results demonstrate that the proposed method outperforms the traditional approaches tested against it while saving considerable computational time.

Classifier ensemble selection may be formulated as a learning task since the search algorithm operates by minimizing/maximizing the objective function. As a consequence, the selection process may be prone to overfitting. The objectives of this paper are: (1) to show how overfitting can be detected when the selection is performed by two classical search algorithms: Genetic Algorithm and Particle Swarm Optimization; and (2) to verify which algorithm is more prone to overfitting. The experimental results demonstrate that GA appears to be more affected by overfitting.

The overproduce-and-choose strategy involves the generation of an initial large pool of candidate classifiers and it is intended to test different candidate ensembles in order to select the best performing solution. The ensemble's error rate, ensemble size and diversity measures are the most frequent search criteria employed to guide this selection. By applying the error rate, we may accomplish the main objective in Pattern Recognition and Machine Learning, which is to find high-performance predictors. In terms of ensemble size, the hope is to increase the recognition rate while minimizing the number of classifiers in order to meet both the performance and low ensemble size requirements. Finally, ensembles can be more accurate than individual classifiers only when classifier members present diversity among themselves. In this paper we apply two Pareto front spread quality measures to analyze the relationship between the three main search criteria used in the overproduce-and-choose strategy. Experimental results conducted demonstrate that the combination of ensemble size and diversity does not produce conflicting multi-objective optimization problems. Moreover, we cannot decrease the generalization error rate by combining this pair of search criteria. However, when the error rate is combined with diversity or the ensemble size, we found that these measures are conflicting objective functions and that the performances of the solutions are much higher.

Defence Research and Development Canada at Valcartier is pursuing the exploration of situation analysis concepts and the prototyping of computer-based decision support systems to maintain the state of situational awareness for the decision maker. The integration of the human element at the beginning of the analysis process is an important facet of our approach. The mathematical formalism and methodology proposed will be illustrated on concrete examples for visibility-based terrain analysis and reasoning for combat search and rescue (CSAR) operations. The work presented is based on the North Atlantis scenario GIS dataset, depicting a conflict taking place over an imaginary continent. The dataset is composed of topographic, hydrographical, transportation, and other typical land cover layers. Applications presented will include landing site determination as well as shortest path to crash site determination.

This paper details and deepens a previous work where the Interpreted Systems semantics was proposed as a general framework for situation analysis (SA). This framework is particularly efficient for representing and reasoning about knowledge and uncertainty when performing situation analysis tasks. Our approach of SA is to base our analysis on the production of state transition systems consisting in the set of all temporal trajectories possibly obtained upon the execution of a given set of agents' protocols. Thus seen, the SA task involves the definition of more or less subtle reasoning about graph structures. A formal situation analysis model is defined as an interpreted algorithmic belief change system. In such a model, the notions of situation, situation awareness and situation analysis are provided. The analysis of the situation is done through the verification of implicit notions of knowledge with temporal properties. Implicit knowledge is distinguished from explicit knowledge and situation awareness is defined in terms of the computing power of resource-bounded agents. A general plausibility measure allows us to model belief while making the link with quantitative representations of uncertainty such as probabilities, belief functions and possibilities. The propsed modelisation of the Situation Analysis process, while compatible with the traditionnal implicit representation of knowledge found in modal logic, allows us to link the decision processes of the agents, their awareness of the situation with the observations they make about the environment.

This paper proposes a discussion on the formal links between the situation calculus and the semantics of interpreted systems as far as they relate to higher-level information fusion tasks. Among these tasks situation analysis require to be able to reason about the decision processes of coalitions. Indeed in higher levels of information fusion, one not only need to know that a certain proposition is true (or that it has a certain numerical measure attached), but rather needs to model the circumstances under which this validity holds as well as agents' properties and constraints. In a previous paper the authors have proposed to use the interpreted system semantics as a potential candidate for the unification of all levels of information fusion. In the present work we show how the proposed framework allow to bind reasoning about courses of action and situation awareness. We propose in this paper a (1) model of coalition, (2) a model of ability in the situation calculus language and (3) a model of situation awareness in the interpreted systems semantics. Combining the advantages of both situation calculus and the interpreted systems semantics, we show how the situation calculus can be framed into the interpreted systems semantics. We illustrate on the example of RAP compilation in a coalition context, how ability and situation awareness interact and what benefit is gained. Finally, we conclude this study with a discussion on possible future works.

The main goal of this paper is to investigate the relationship between two theories widely applied to explain the success of classifiers fusion: diversity measures and margin theory. In order to achieve this, we realized an empirical study which evaluates some classical measures related to these two theories with respect to ensembles accuracy. In particular, this study revealed valuable insights on how these two theories can influence each other, and how the application of margin based measures can be useful for the evaluation and selection of ensembles of classifiers with majority voting.

Dynamic classifier selection has traditionally focused on selecting the most accurate classifier to predict the class of a particular test pattern. In this paper we propose a new dynamic selection method to select, from a population of ensembles, the most confident ensemble of classifiers to label the test sample. Such a level of confidence is measured by calculating the ambiguity of the ensemble on each test sample. We show theoretically and experimentally that choosing the ensemble of classifiers, from a population of high accurate ensembles, with lowest ambiguity among its members leads to increase the level of confidence of classification, consequently, increasing the generalization performance. Experimental results conducted to compare the proposed method to static selection and DCS-LA, demonstrate that our method outperforms both DCS-LA and static selection strategies when a population of high accurate ensembles is available.

Many recent works have investigated methods to select subsets of classifiers instead of combining all available classifiers. The majority of these works has concluded that the combiner error rate is better than diversity to guide the selection process in order to identify the best performing subset of classifiers. However, the classifier selection process has to take into account three different aspects: complexity, overfitting and performance. These aspects of the selection process have not yet been tackled simultaneously in the literature. The study presented in this paper, deals with these three aspects in a handwritten digit recognition problem. Different search criteria such as diversity, error rate and number of classifiers are applied in single and multi-objective optimization approaches using genetic algorithms. In our experiments, we observed that error rate applied in a single optimization approach was the best objective function to increase performance. The generalized diversity and interrater agreement measures, combined with error rate in pairs of objective functions were the best measures to reduce complexity and keep good performance in a multi-objective optimization approach. Finally, the performance of the solutions found in both, single and multi-objective optimization processes were increased by applying a global validation method to reduce overfitting.

The aim of this paper is to present the results of a study on the relationship between the presence ofAmbrosia artemisiifolia(Common ragweed) and the characteristics of physical and social environments of Montreal City. Standard multiple regression techniques allowed to create several models with good fitting values (R2>0.65) and statistically significant using the density of observations of Ambrosia artemisiifolia as the dependent variable. The present study, realized at the city block level, showed that the joint use of remote sensing imagery, land-use databases and socio-economic census data is very promising for the predictive cartography ofAmbrosia artemisiifolia. To the authors knowledge it is the first such modeling attempt published yet

This papers proposes a reflexion on the concept of vagueness. We relate the different kinds of vagueness to relevant mathematical frameworks with particular emphasis on recent works on neutrosophic logic, Dezert-Smarandache theory, and rough sets. The different facets of vagueness are illustrated through a case study on Ambrosia artemisiifolia predictive cartography