International Joint Conference on Neural Networks

Rockburst prediction is of crucial importance to the design and construction of many underground projects. Insufficient knowledge, lack of characterizing information, and noisy data restrain rock mechanics engineers from achieving optimal prediction results. In this paper, a data set of 246 rockburst events was examined for rockburst classification using supervised learning (SL) methods. The data set was analyzed with 8 potentially relevant indicators. Eleven algorithms from 10 categories of SL algorithms were evaluated for their ability to learn rockburst, including linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), partial least-squares discriminant analysis (PLSDA), naive Bayes (NB), k-nearest neighbor (KNN), multilayer perceptron neural network (MLPNN), classification tree (CT), support vector machine (SVM), random forest (RF), and gradient-boosting machine (GBM). The data set was randomly split into two parts: training (70%) and test (30%). A 10-fold cross-validation (...

Classification of Rockburst in Underground Projects: Comparison of Ten Supervised Learning Methods

Missing value imputation (MVI) has been studied for several decades being the basic solution method for incomplete dataset problems, specifically those where some data samples contain one or more missing attribute values. This paper aims at reviewing and analyzing related studies carried out in recent decades, from the experimental design perspective. Altogether, 111 journal papers published from 2006 to 2017 are reviewed and analyzed. In addition, several technical issues encountered during the MVI process are addressed, such as the choice of datasets, missing rates and missingness mechanisms, and the MVI techniques and evaluation metrics employed, are discussed. The results of analysis of these issues allow limitations in the existing body of literature to be identified based upon which some directions for future research can be gleaned.

Missing value imputation: a review and analysis of the literature (2006–2017)

The world's oceans are undergoing profound changes as a result of human activities. However, the consequences of escalating human impacts on marine mammal biodiversity remain poorly understood. The International Union for the Conservation of Nature (IUCN) identifies 25% of marine mammals as at risk of extinction, but the conservation status of nearly 40% of marine mammals remains unknown due to insufficient data. Predictive models of extinction risk are crucial to informing present and future conservation needs, yet such models have not been developed for marine mammals. In this paper, we: (i) used powerful machine-learning and spatial-modeling approaches to understand the intrinsic and extrinsic drivers of marine mammal extinction risk; (ii) used this information to predict risk across all marine mammals, including IUCN "Data Deficient" species; and (iii) conducted a spatially explicit assessment of these results to understand how risk is distributed across the world's oceans. Rate of offspring production was the most important predictor of risk. Additional predictors included taxonomic group, small geographic range area, and small social group size. Although the interaction of both intrinsic and extrinsic variables was important in predicting risk, overall, intrinsic traits were more important than extrinsic variables. In addition to the 32 species already on the IUCN Red List, our model identified 15 more species, suggesting that 37% of all marine mammals are at risk of extinction. Most at-risk species occur in coastal areas and in productive regions of the high seas. We identify 13 global hotspots of risk and show how they overlap with human impacts and Marine Protected Areas.

/pdf/drivers-and-hotspots-of-extinction-risk-in-marine-mammals-165owsk4cf.pdf

Drivers and hotspots of extinction risk in marine mammals

This paper provides insight into the use of Machine Learning (ML) models for the assessment of human-caused wildfire occurrence. It proposes the use of ML within the context of fire risk prediction, and more specifically, in the evaluation of human-induced wildfires in Spain. In this context, three ML algorithms—Random Forest (RF), Boosting Regression Trees (BRT), and Support Vector Machines (SVM)—are implemented and compared with traditional methods like Logistic Regression (LR). Results suggest that the use of any of these ML algorithms leads to an improvement in the accuracy—in terms of the AUC (area under the curve)—of the model when compared to LR outputs. According to the AUC values, RF and BRT seem to be the most adequate methods, reaching AUC values of 0.746 and 0.730 respectively. On the other hand, despite the fact that the SVM yields an AUC value higher than that from LR, the authors consider it inadequate for classifying wildfire occurrences because its calibration is extremely time-consuming.

An insight into machine-learning algorithms to model human-caused wildfire occurrence

Missing value imputation on missing completely at random data using multilayer perceptrons

Oak forests are essential for the ecosystems of many countries, particularly when they are used in vegetal restoration. Therefore, models for predicting the potential habitat of oaks can be a valuable tool for work in the environment. In accordance with this objective, the building and comparison of data mining models are presented for the prediction of potential habitats for the oak forest type in Mediterranean areas (southern Spain), with conclusions applicable to other regions. Thirty-one environmental input variables were measured and six base models for supervised classification problems were selected: linear and quadratic discriminant analysis, logistic regression, classification trees, neural networks and support vector machines. Three ensemble methods, based on the combination of classification tree models fitted from samples and sets of variables generated from the original data set were also evaluated: bagging, random forests and boosting. The available data set was randomly split into three parts: training set (50%), validation set (25%), and test set (25%). The analysis of the accuracy, the sensitivity, the specificity, together with the area under the ROC curve for the test set reveal that the best models for our oak data set are those of bagging and random forests. All of these models can be fitted by free R programs which use the libraries and functions described in this paper. Furthermore, the methodology used in this study will allow researchers to determine the potential distribution of oaks in other kinds of areas.

Predicting the potential habitat of oaks with data mining models and the R system

A wide range of supervised classification algorithms have been successfully applied for credit scoring in non-microfinance environments according to recent literature. However, credit scoring in the microfinance industry is a relatively recent application, and current research is based, to the best of our knowledge, on classical statistical methods. This lack is surprising since the implementation of credit scoring based on supervised classification algorithms should contribute towards the efficiency of microfinance institutions, thereby improving their competitiveness in an increasingly constrained environment. This paper explores an extensive list of Statistical Learning techniques as microfinance credit scoring tools from an empirical viewpoint. A data set of microcredits belonging to a Peruvian Microfinance Institution is considered, and the following models are applied to decide between default and non-default credits: linear and quadratic discriminant analysis, logistic regression, multilayer perceptron, support vector machines, classification trees, and ensemble methods based on bagging and boosting algorithm. The obtained results suggest the use of a multilayer perceptron trained in the R statistical system with a second order algorithm. Moreover, our findings show that, with the implementation of this MLP-based model, the MFIs@? misclassification costs could be reduced to 13.7% with respect to the application of other classic models.

Improving the management of microfinance institutions by using credit scoring models based on Statistical Learning techniques

Reduced bootstrap aggregating of learning algorithms

Summary We define a variation of Efron's method II based on the outlier bootstrap sample concept. A criterion for the identification of such samples is given, with which a variation in the bootstrap sample generation algorithm is introduced. The results of several simulations are analyzed in which, in comparison with Efron's method II, a higher degree of closeness to the estimated quantities can be observed.

María-Dolores Cubiles-de-la-Vega

Papers

Missing value imputation on missing completely at random data using multilayer perceptrons

Predicting the potential habitat of oaks with data mining models and the R system

Improving the management of microfinance institutions by using credit scoring models based on Statistical Learning techniques

Reduced bootstrap aggregating of learning algorithms

Identification of outlier bootstrap samples