Search

In poultry houses, thermal comfort is decisive for maximising the feed conversion rate, a measure for successful production. As there are different automatic control options, that result in variable performance indexes, this paper reproduces, tests, and compares two of them: a reactive, that applies event-driven methods, and a bio-inspired, which is based on artificial intelligence techniques. As each approach adds specific advantages to the process control, we combine them into a single framework that gathers their best features. Simulations using real data show that temperature and humidity have been reproduced with 97% and 80% precision, respectively. In comparison, reactive and bio-inspired approaches show respectively accuracies of 90% and 82% for temperature, and 60% and 66% for humidity. Therefore, we conclude that our approach can improve both reactive and bio-inspired control, standing as a feasible and flexible alternative for the control of poultry houses.

Supervised classification is a mechanism used in machine learning to associate classes with objects from datasets. Depending on the dimension and on the internal data structuring, classification may become complex. In this paper, we claim that the complexity level of a given dataset can be estimated by using fractal analysis. A novel fractal measure, called transition border, is proposed in order to estimate the chaos behind labelled points distribution. Their correlation with the success rateis tested by comparing it against results obtained from other supervised classification methods. Results suggest that this approach can be used to measure the complexity behind a classification task problem in real-valued datasets with three dimensions. The proposed method can also be useful for other science domains for which fractal analysis is applicable.

Abstract: In this paper, we propose a novel texture analysis method using the complex network theory. We investigated how a texture image can be effectively represented, characterized and analyzed in terms of a complex network. The proposed approach uses degree measurements to compose a set of texture descriptors. The results show that the method is very robust, and it presents a excellent texture discrimination for all considered classes, overcoming traditional texture methods. [Copyright &y& Elsevier]

This work proposes a method to extract features from texture images by applying a Gaussian pyramid multiscale approach to the Bouligand–Minkowski fractal descriptors. The proposal starts from the texture image and computes the stack of multi-resolution images that compose the pyramid, in both directions, of reduction and expansion. In the following, each image in the stack is mapped onto a surface, which is dilated by spheres with variable radii and the dilation volumes are used to compute the Bouligand–Minkowski fractal descriptors for each level. Both the descriptors of each level and combinations with descriptors from the original image are verified in the classification of well-known databases of textural images. The proposed method outperformed other classical and state-of-the-art descriptors with a significant advantage in most cases, including situations where random noise is added to the images. [ABSTRACT FROM AUTHOR]

• Bouligand-Minkowski is based on the dilation of a shape with radius r. • To use it on networks, edges are interpreted as pipes giving vent to some liquid. • Vertices are end pipes and the dilation is the insertion of fluids in the system. • Fluids are inserted into the edges by the vertex and shared with adjacent edges. A complex network presents many topological features which characterize its behavior and dynamics. This characterization is an essential aspect of complex networks analysis and can be performed using several measures, including the fractal dimension. Originally the fractal dimension measures the complexity of an object in a Euclidean space, and the most common methods in the literature to estimate that dimension are box-counting, mass-radius, and Bouligand-Minkowski. However, networks are not Euclidean objects, so that these methods require some adaptation to measure the fractal dimension in this context. The literature presents some adaptations for methods like box-counting and mass-radius. However, there is no known adaptation developed for the Bouligand-Minkowski method. In this way, we propose an adaptation of the Bouligand-Minkowski to measure complex networks' fractal dimension. We compare our proposed method with others, and we also explore the application of the proposed method in a classification task of complex networks that confirmed its promising potential. [ABSTRACT FROM AUTHOR]