Search

Purpose This study aims to perform a benefit segmentation and then a classification of visitors that travel to the Rocha Department in Uruguay from the capital city of Montevideo during the summer months. Design/methodology/approach A convenience sample was obtained with an online survey. A total of 290 cases were usable for subsequent data analysis. The following statistical techniques were used: hierarchical cluster analysis, K-means cluster analysis, machine learning, support vector machines, random forest and logistic regression. Findings Visitors that travel to the Rocha Department from Montevideo can be classified into four distinct clusters. Clusters are labelled as “entertainment seekers”, “Rocha followers”, “relax and activities seekers” and “active tourists”. The support vector machine model achieved the best classification results. Research limitations/implications Implications for destination marketers who cater to young visitors are discussed. Destination marketers should determine an optimal level of resource allocation and destination management activities that compare both present costs and discounted potential future income of the different target markets. Surveying non-residents was not possible. Future work should sample tourists from abroad. Originality/value The combination of market segmentation of Rocha Department’s visitors from the city of Montevideo and classification of sampled individuals training various machine learning classifiers would allow Rocha’s destination marketers determine the belonging of an unsampled individual into one of the already obtained four clusters, enhancing marketing promotion for targeted offers.

En este trabajo se construyeron modelos de predicción de coliformes fecales (CF) para su aplicación a la gestión de calidad de agua en playas recreativas. Se análizó la base de datos histórica del monitoreo de CF en playas realizado por el Laboratorio de Calidad Ambiental de la Intendencia de Montevideo (IM), y se generaron modelos basados en inteligencia artificial (IA) para predecir excesos (CF >2.000). Los datos abarcan 10 años de monitoreo en 21 playas de la capital (N=19359, noviembre 2009 a setiembre 2019) y presentan un amplio rango de salinidad y turbidez, con marcadas diferencias entre playas. Los CF presentaron una distribución asimétrica (min=4, mediana=250, media=1.047 y máx=1.280.000) con excesos a la normativa en todas las playas. Las variables registradas in situ, meteorológicas y oceanográficas fueron utilizadas para entrenar modelos de IA. El mejor modelo fue un bosque aleatorio estratificado con un porcentaje de acierto para los excesos de 86%. La predicción de excesos mejoró un 60% respecto al criterio actual de cierre de playas las 24 hs posteriores a precipitaciones. La generación de datos de calidad por parte de la IM junto con estrategias de modelización inteligente resultan en un insumo relevante para la gestión de playas recreativas.

In this work we give new density estimators by averaging classical density estimators such as the histogram, the frequency polygon and the kernel density estimators obtained over different bootstrap samples of the original data. We prove the L 2-consistency of these new estimators and compare them to several similar approaches by extensive simulations. Based on them, we give also a way to construct non parametric pointwise confidence intervals for the target density.

Predicting water contamination by statistical models is a useful tool to manage health risk in recreational beaches. Extreme contamination events, i.e. those exceeding normative are generally rare with respect to bathing conditions and thus the data is said to be imbalanced. Modeling and predicting those rare events present unique challenges. Here we introduce and evaluate several machine learning techniques and metrics to model imbalanced data and evaluate model performance. We do so by using a) simulated data-sets and b) a real data base with records of faecal coliform abundance monitored for 10 years in 21 recreational beaches in Uruguay (N ≈ 19000) using in situ and meteorological variables. We discuss advantages and disadvantages of the methods and provide a simple guide to perform models for a general audience. We also provide R codes to reproduce model fitting and testing. We found that most Machine Learning techniques are sensitive to imbalance and require specific data pre-treatment (e.g. upsampling) to improve performance. Accuracy (i.e. correctly classified cases over total cases) is not adequate to evaluate model performance on imbalanced data set. Instead, true positive rates (TPR) and false positive rates (FPR) are recommended. Among the 52 possible candidate algorithms tested, the stratified Random forest presented the better performance improving TPR in 50% with respect to baseline (0.4) and outperformed baseline in the evaluated metrics. Support vector machines combined with upsampling method or synthetic minority oversampling technique (SMOTE) performed well, similar to Adaboost with SMOTE. These results suggests that combining modeling strategies is necessary to improve our capacity to anticipate water contamination and avoid health risk.

Multiclass classification refers to the construction of a model able to classify a response variable that can take more than two classes. Most ecological indices are naturally multiclass (e.g. water quality index: bad, regular, good) and the generation of models able to predict the output class in novel situations is required. In this study, we introduce seven representative multiclass classification techniques, classic and more recent, their rationale, advantages, disadvantages and a practical R code to implement them. These methods are: (1) Linear discriminant analysis (LDA), (2) a consensus of binomial logistic regression (CLR), (3) multinomial regression (MNR) and (4) support vector machine (SVM), (5) Classification and Regression Trees (CART), (6) Random Forest (RF) and (7) Stagewise Additive Modelling using a Multi-class Exponential (SAMME) loss function. We showed their implementation under simulated and a real data set to classify phytoplankton organisms into morphology-based functional groups. Results suggest that the nature of the data (i.e. linear vs non-linear) influence the predictive ability of multi-class classification models. Real phytoplankton data was accurately classified (error

We performed different consensus methods by combining binary classifiers, mostly machine learning classifiers, with the aim to test their capability as predictive tools for the presence–absence of marine phytoplankton species. The consensus methods were constructed by considering a combination of four methods (i.e., generalized linear models, random forests, boosting and support vector machines). Six different consensus methods were analyzed by taking into account six different ways of combining single-model predictions. Some of these methods are presented here for the first time. To evaluate the performance of the models, we considered eight phytoplankton species presence–absence data sets and data related to environmental variables. Some of the analyzed species are toxic, whereas others provoke water discoloration, which can cause alarm in the population. Besides the phytoplankton data sets, we tested the models on 10 well-known open access data sets. We evaluated the models' performances over a test sample. For most (72%) of the data sets, a consensus method was the method with the lowest classification error. In particular, a consensus method that weighted single-model predictions in accordance with single-model performances (weighted average prediction error — WA-PE model) was the one that presented the lowest classification error most of the time. For the phytoplankton species, the errors of the WA-PE model were between 10% for the species Akashiwo sanguinea and 38% for Dinophysis acuminata . This study provides novel approaches to improve the prediction accuracy in species distribution studies and, in particular, in those concerning marine phytoplankton species.

There is a strong interplay between network reliability and connectivity theory. In fact, previous studies show that the graphs with maximum reliability, called uniformly most-reliable graphs, must have the highest connectivity. In this paper, we revisit the underlying theory in order to build uniformly most-reliable cubic graphs. The computational complexity of the problem promotes the development of heuristics. The contributions of this paper are three-fold. In a first stage, we propose an ideal Variable Neighborhood Descent (VND) which returns the graph with maximum reliability. This VND works in exponential time. In a second stage, we propose a Greedy Randomized Adaptive Search Procedure (GRASP), that trades quality for computational effort. A construction phase enriched with a Restricted Candidate List (RCL) offers diversification. Our local search phase includes a globally optimum solution of an Integer Linear Programming (ILP) formulation. As a product of our research, we recovered previous optimal graphs from the related literature in the field. Additionally, we offer new candidates of uniformly most-reliable graphs with maximum connectivity and maximum number of spanning trees.

In this paper we address a metaheuristic for an combinatorial optimization problem. For any given graph \(\mathcal {G}=(V,E)\) (where the nodes represent items and edges correlations), we want to find the clique \(\mathcal {C} \subseteq V\) such that the number of links shared between \(\mathcal {C}\) and \(V - \mathcal {C}\) is maximized. This problem is known in the literature as the Max Cut-Clique (MCC).

In Market Basket Analysis, the goal is to understand the human behavior in order to maximize sales. An evident behavior is to buy correlated items. As a consequence, the determination of a set of items with a large correlation with others is a valuable tool for Market Basket Analysis.

There is a strong interplay between network reliability and connectivity theory. In fact, previous studies show that the graphs with maximum reliability, called uniformly most-reliable graphs, must have the highest connectivity. In this paper, we revisit the underlying theory in order to build uniformly most-reliable cubic graphs. The computational complexity of the problem promotes the development of heuristics. The contributions of this paper are three-fold. In a first stage, we propose an ideal Variable Neighborhood Descent (VND) which returns the graph with maximum reliability. This VND works in exponential time. In a second stage, we propose a hybrid GRASP/VND approach that trades quality for computational effort. A construction phase enriched with a Restricted Candidate List (RCL) offers diversification. Our local search phase includes a factor-2 algorithm for an Integer Linear Programming (ILP) model. As a product of our research, we recovered previous optimal graphs from the related literature in the field. Additionally, we offer new candidates of uniformly most-reliable graphs with maximum connectivity and maximum number of spanning trees.

International audience; We present a new multiclass boosting algorithm called Adaboost.BG. Like the original Freund and Shapire's Adaboost algorithm, it aggregates trees but instead of using their misclassification error it takes into account the margins of the observations, which may be seen as confidence measures of their prediction, rather then their correctness. We prove the efficiency of our algorithm by simulation and compare it to similar approaches known to minimize the global margins of the final classifier.

International audience; A new density estimator called RASH, for Random Average Shifted Histogram, obtained by averaging several histograms as proposed in average shifted histograms, is presented. The principal difference between the two methods is that in RASH each histogram is built over a grid with random shifted breakpoints. The asymptotic behavior of this estimator is established for the one-dimensional case and its performance through several simulations is analyzed. RASH is compared to several classic density estimators and to some recent ensemble methods. Although RASH does not always outperform the other methods, it is very simple to implement, being also more intuitive. The two dimensional case is also analyzed empirically.

Let T be a torus over an algebraically closed field k of characteristic 0, and consider a projective T-module P(V). We determine when a projective toric subvariety X of P(V) is self-dual, in terms of the configuration of weights of V., Comment: 26 pages, 1 figure. Minor changes

We present some new density estimation algorithms obtained by bootstrap aggregation like Bagging. Our algorithms are analyzed and empirically compared to other methods found in the statistical literature, like stacking and boosting for density estimation. We show by extensive simulations that ensemble learning are effective for density estimation like for classification. Although our algorithms do not always outperform other methods, some of them are as simple as bagging, more intuitive and has computational lower cost.