A study on the use of statistical tests for experimentation with neural networks: Analysis of parametric test conditions and non-parametric tests

In this paper, we focus on the experimental analysis on the performance in artificial neural networks with the use of statistical tests on the classification task. Particularly, we have studied whether the sample of results from multiple trials obtained by conventional artificial neural networks and support vector machines checks the necessary conditions for being analyzed through parametrical tests. The study is conducted by considering three possibilities on classification experiments: random variation in the selection of test data, the selection of training data and internal randomness in the learning algorithm.The results obtained state that the fulfillment of these conditions are problem-dependent and indefinite, which justifies the need of using non-parametric statistics in the experimental analysis.     

Automatic Schaeffer's gestures recognition system

Schaeffer's sign language consists of a reduced set of gestures designed to help children with autism or cognitive learning disabilities to develop adequate communication skills. Our automatic recognition system for Schaeffer's gesture language uses the information provided by an RGB‐D camera to capture body motion and recognize gestures using dynamic time warping combined with k‐nearest neighbors methods. The learning process is reinforced by the interaction with the proposed system that accelerates learning itself thus helping both children and educators. To demonstrate the validity of the system, a set of qualitative experiments with children were carried out. As a result, a system which is able to recognize a subset of 11 gestures of Schaeffer's sign language online was achieved.     

Ordinal regression by a gravitational model in the field of educational data mining

Educational data mining (EDM) is a research area where the goal is to develop data mining methods to examine data critically from educational environments. Traditionally, EDM has addressed the following problems: clustering, classification, regression, anomaly detection and association rule mining. In this paper, the ordinal regression (OR) paradigm, is introduced in the field of EDM. The goal of OR problems is the classification of items in an ordinal scale. For instance, the prediction of students' performance in categories (where the different grades could be ordered according to A ? B ? C ? D) is a classical example of an OR problem. The EDM community has not yet explored this paradigm (despite the importance of these problems in the field of EDM). Furthermore, an amenable and interpretable OR model based on the concept of gravitation is proposed. The model is an extension of a recently proposed gravitational model that tackles imbalanced nominal classification problems. The model is carefully adapted to the ordinal scenario and validated with four EDM datasets. The results obtained were compared with state‐of‐the‐art OR algorithms and nominal classification ones. The proposed models can be used to better understand the learning–teaching process in higher education environments.     

A case study about the improvement of business process models driven by indicators

Organizations are increasingly concerned about business process model improvement in their efforts to guarantee improved operational efficiency. Quality assurance of business process models should be addressed in the most objective manner, e.g., through the application of measures, but the assessment of measurement results is not a straightforward task and it requires the identification of relevant indicators and threshold values, which are able to distinguish different levels of process model quality. Furthermore, indicators must support the improvements of the models by using suitable guidelines. In this paper, we present a case study to evaluate the BPMIMA framework for BP model improvement. This framework is composed of empirically validated measures related to quality characteristics of the models, a set of indicators with validated thresholds associated with modeling guidelines and a prototype supporting tool. The obtained data suggest that the redesign by applying guidelines driven by the indicator results was successful, as the understandability and modifiability of the models were improved. In addition, the changes in the models according to guidelines were perceived as acceptable by the practitioners who participated in the case study.     

Decentralized MDPs with sparse interactions

Creating coordinated multiagent policies in environments with uncertainty is a challenging problem, which can be greatly simplified if the coordination needs are known to be limited to specific parts of the state space. In this work, we explore how such local interactions can simplify coordination in multiagent systems. We focus on problems in which the interaction between the agents is sparse and contribute a new decision-theoretic model for decentralized sparse-interaction multiagent systems, Dec-SIMDPs, that explicitly distinguishes the situations in which the agents in the team must coordinate from those in which they can act independently. We relate our new model to other existing models such as MMDPs and Dec-MDPs. We then propose a solution method that takes advantage of the particular structure of Dec-SIMDPs and provide theoretical error bounds on the quality of the obtained solution. Finally, we show a reinforcement learning algorithm in which independent agents learn both individual policies and when and how to coordinate. We illustrate the application of the algorithms throughout the paper in several multiagent navigation scenarios.     

A dynamic over-sampling procedure based on sensitivity for multi-class problems

Classification with imbalanced datasets supposes a new challenge for researches in the framework of machine learning. This problem appears when the number of patterns that represents one of the classes of the dataset (usually the concept of interest) is much lower than in the remaining classes. Thus, the learning model must be adapted to this situation, which is very common in real applications. In this paper, a dynamic over-sampling procedure is proposed for improving the classification of imbalanced datasets with more than two classes. This procedure is incorporated into a memetic algorithm (MA) that optimizes radial basis functions neural networks (RBFNNs). To handle class imbalance, the training data are resampled in two stages. In the first stage, an over-sampling procedure is applied to the minority class to balance in part the size of the classes. Then, the MA is run and the data are over-sampled in different generations of the evolution, generating new patterns of the minimum sensitivity class (the class with the worst accuracy for the best RBFNN of the population). The methodology proposed is tested using 13 imbalanced benchmark classification datasets from well-known machine learning problems and one complex problem of microbial growth. It is compared to other neural network methods specifically designed for handling imbalanced data. These methods include different over-sampling procedures in the preprocessing stage, a threshold-moving method where the output threshold is moved toward inexpensive classes and ensembles approaches combining the models obtained with these techniques. The results show that our proposal is able to improve the sensitivity in the generalization set and obtains both a high accuracy level and a good classification level for each class.     

The complementary exponential geometric distribution: Model, properties, and a comparison with its counterpart

In this paper, we proposed a new two-parameter lifetime distribution with increasing failure rate, the complementary exponential geometric distribution, which is complementary to the exponential geometric model proposed by Adamidis and Loukas (1998). The new distribution arises on a latent complementary risks scenario, in which the lifetime associated with a particular risk is not observable; rather, we observe only the maximum lifetime value among all risks. The properties of the proposed distribution are discussed, including a formal proof of its probability density function and explicit algebraic formulas for its reliability and failure rate functions, moments, including the mean and variance, variation coefficient, and modal value. The parameter estimation is based on the usual maximum likelihood approach. We report the results of a misspecification simulation study performed in order to assess the extent of misspecification errors when testing the exponential geometric distribution against our complementary one in the presence of different sample size and censoring percentage. The methodology is illustrated on four real datasets; we also make a comparison between both modeling approaches.     

Proof Pearl: a Formal Proof of Higman��s Lemma in ACL2

Higman??s lemma is an important result in infinitary combinatorics, which has been formalized in several theorem provers. In this paper we present a formalization and proof of Higman??s Lemma in the ACL2 theorem prover. Our formalization is based on a proof by Murthy and Russell, where the key termination argument is justified by the multiset relation induced by a well-founded relation. To our knowledge, this is the first mechanization of this proof.     

Illustrating volume data sets and layered models with importance-aware composition

Non-photorealistic (illustrative) rendering augments typical rendering models to selectively emphasize or de-emphasize specific structures of rendered objects. Illustrative techniques may affect not only the rendering style of specific portions of an object but also their visibility, ensuring that less important regions do not occlude more important ones. Cutaway views completely remove occluding, unimportant structures—possibly also removing valuable context information—while existing solutions for smooth reduction of occlusion based on importance lack precise visibility control, simplicity and generality. We introduce a new front-to-back fragment composition equation that directly takes into account a measure of sample importance and allows smooth and precise importance-based visibility control. We demonstrate the generality of our composition equation with several illustrative effects, which were obtained by using a set of importance measures calculated on the fly or defined by the user. The presented composition method is suitable for direct volume rendering as well as rendering of layered 3D models. We discuss both cases and show examples, though focusing mainly on illustration of volumetric data.     

Visualization of level-of-detail meshes on the GPU

Extensive research has been carried out in multiresolution models for many decades. The tendency in recent years has been to harness the potential of GPUs to perform the level-of-detail extraction on graphics hardware. The aim of this work is to present a new level-of-detail scheme based on triangles which is both simple and efficient. In this approach, the extraction process updates vertices instead of indices, thus providing a perfect framework for adapting the algorithms to work completely on GPU shaders. One of the key aspects of our proposal is the need for just a single rendering pass in order to obtain the desired geometry. Moreover, coherence among the different approximations is maximized by means of a symmetric extraction algorithm, which performs the same process when refining and coarsening the mesh. Lastly, we also introduce different uses of the scheme to offer continuous and view-dependent resolution.