Unravelling migratory connectivity in marine turtles using multiple methods

1. Comprehensive knowledge of the fundamental spatial ecology of marine species is critical to allow the identification of key habitats and the likely sources of anthropogenic threats, thus informing effective conservation strategies. 2. Research on migratory marine vertebrates has lagged behind many similar terrestrial animal groups, but studies using electronic tagging systems and molecular techniques offer great insights. 3. Marine turtles have complex life history patterns, spanning wide spatio-temporal scales. As a result of this multidimensional complexity, and despite extensive effort, there are no populations for which a truly holistic understanding of the spatial aspects of the life history has been attained. There is a particular lack of information regarding the distribution and habitats utilized during the first few years of life. 4. We used satellite tracking technology to track individual turtles following nesting at the green turtle Chelonia mydas nesting colony at Poilão Island, Guinea Bissau; the largest breeding aggregation in the eastern Atlantic. 5. We further contextualize these data with pan-Atlantic molecular data and oceanographic current modelling to gain insights into likely dispersal patterns of hatchlings and small pelagic juveniles. 6. All adult turtles remained in the waters of West Africa, with strong connectivity demonstrated with Banc D’Arguin, Mauritania. 7. Despite shortcomings in current molecular markers, we demonstrate evidence for profound sub-structuring of marine turtle stocks across the Atlantic; with a high likelihood based on oceanographic modelling that most turtles from Guinea-Bissau are found in the eastern Atlantic. 8. Synthesis and applications. There is an increased need for a better understanding of spatial distribution of marine vertebrates demonstrating life histories with spatio-temporal complexity. We propose the synergistic use of the technologies and modelling used here as a working framework for the future rapid elucidation of the range and likely key habitats used by the different life stages from such species.

Image reduction operators as aggregation functions: Fuzzy transform and undersampling

After studying several reduction algorithms that can be found in the literature, we notice that there is not an axiomatic definition of this concept. In this work we propose the definition of weak reduction operators and we propose the properties of the original image that reduced images must keep. From this definition, we study whether two methods of image reduction, undersampling and fuzzy transform, satisfy the conditions of weak reduction operators.

Examining performance of aggregation algorithms for neural network-based electricity demand forecasting

The aim of this research is to examine the efficiency of different aggregation algorithms to the forecasts obtained from individual neural network (NN) models in an ensemble. In this study an ensemble of 100 NN models are constructed with a heterogeneous architecture. The outputs from NN models are combined by three different aggregation algorithms. These aggregation algorithms comprise of a simple average, trimmed mean, and a Bayesian model averaging. These methods are utilized with certain modifications and are employed on the forecasts obtained from all individual NN models. The output of the aggregation algorithms is analyzed and compared with the individual NN models used in NN ensemble and with a Naive approach. Thirty-minutes interval electricity demand data from Australian Energy Market Operator (AEMO) and the New York Independent System Operator's web site (NYISO) are used in the empirical analysis. It is observed that the aggregation algorithm perform better than many of the individual NN models. In comparison with the Naive approach, the aggregation algorithms exhibit somewhat better forecasting performance.

Aggregation functions for recommender systems

This chapter gives an overview of aggregation functions and their use in recommender systems. The classical weighted average lies at the heart of various recommendation mechanisms, often being employed to combine item feature scores or predict ratings from similar users. Some improvements to accuracy and robustness can be achieved by aggregating different measures of similarity or using an average of recommendations obtained through different techniques. Advances made in the theory of aggregation functions therefore have the potential to deliver increased performance to many recommender systems. We provide definitions of some important families and properties, sophisticated methods of construction, and various examples of aggregation functions in the domain of recommender systems.

Linear optimization for ecological indices based on aggregation functions

We consider an optimization problem in ecology where our objective is to maximize biodiversity with respect to different land-use allocations. As it turns out, the main problem can be framed as learning the weights of a weighted arithmetic mean where the objective is the geometric mean of its outputs. We propose methods for approximating solutions to this and similar problems, which are non-linear by nature, using linear and bilevel techniques.

Fitting aggregation functions to data: part I-linearization and regularization

The use of supervised learning techniques for fitting weights and/or generator functions of weighted quasi-arithmetic means – a special class of idempotent and nondecreasing aggregation functions – to empirical data has already been considered in a number of papers. Nevertheless, there are still some important issues that have not been discussed in the literature yet. In the first part of this two-part contribution we deal with the concept of regularization, a quite standard technique from machine learning applied so as to increase the fit quality on test and validation data samples. Due to the constraints on the weighting vector, it turns out that quite different methods can be used in the current framework, as compared to regression models. Moreover, it is worth noting that so far fitting weighted quasi-arithmetic means to empirical data has only been performed approximately, via the so-called linearization technique. In this paper we consider exact solutions to such special optimization tasks and indicate cases where linearization leads to much worse solutions.

Aggregation and consensus for preference relations based on fuzzy partial orders

We propose a framework for eliciting and aggregating pairwise preference relations based on the assumption of an underlying fuzzy partial order. We also propose some linear programming optimization methods for ensuring consistency either as part of the aggregation phase or as a pre- or post-processing task. We contend that this framework of pairwise-preference relations, based on the Kemeny distance, can be less sensitive to extreme or biased opinions and is also less complex to elicit from experts. We provide some examples and outline their relevant properties and associated concepts.