Scipedia: David Gabriel de Barros Franco's personal collection

New Strategies for Initialization and Training of Radial Basis Function Neural Networks

David Gabriel de Barros Franco — Sun, 20 Mar 2022 17:01:03 +0100

In this paper we proposed two new strategies for initialization and training of Radial Basis Function (RBF) Neural Network. The first approach takes into consideration the "error" between the input vector p of the network and the x-axis, which are the centers of radial functions. The second approach takes into account the "error" between the input vector p and the network output. In order to check the performances of these strategies, we used Brazilian financial market data for the RBF networks training, specifically the adjusted prices of the 10 greater weighted shares in the Bovespa index at the time of data collection - from April 8th , 2009 to October 31th , 2014. The first approach presented a 52% of improvement in the mean squared error (MSE) compared to the standard RBF network, while the improvement for the second approach was 38%. The strategies proved to be consistent for the time series tested, in addition to having a low computational cost. It is proposed that these strategies be improved by testing them with the Levenberg-Marquardt algorithm.

Classification Algorithms in Financial Application: Credit Risk Analysis on Legal Entities

David Gabriel de Barros Franco — Sun, 20 Mar 2022 16:55:02 +0100

This research aims at analyzing bank credit of legal entity (in non-default, default and temporarily default), for the purpose of assisting the decision made by the analyst of this area. For that, we used Artificial Neural Networks (ANNs), more specifically, the Multilayer Perceptron (MLP) and the Radial Basis Functions (RBF) and, also, the statistical model of Logistic Regression (LR). For the implementation of the ANNs and LR, the softwares MATLAB and SPSS were used, respectively. For the simulations developed 5.432 data with 15 attributes were collected by the experts of the institution bank (called “XYZ”). The results show that the default clients are easily identifiable, but for the nondelinquent clients and for the temporarily defaulters, the techniques had greater difficulty in the discrimination, suggesting that they are no so discriminants. The main contributions of this work are: the analysis of three classes of clients (non-default, default and temporarily default), rather than just two (non-default and default) as is usually done; the coding of variables (attributes) of the company XYZ aiming to maximize the accuracy of the techniques and the use of the one-against all method, little used by the researchers of this research area. This work presents new insights towards research over Credit Risk Assessment showing other possibilities of client classification and codification, allowing different types of studies to take place.

Classification of Abandoned Areas for Solar Energy Projects Using Artificial Intelligence and Quantum Mechanics

David Gabriel de Barros Franco — Sun, 20 Mar 2022 16:47:03 +0100

The increasing demand for energy has intensified recently, requiring alternative sources to fossil fuels, which have become economically and environmentally unfeasible. On the other hand, the increasing land occupation in recent centuries is a growing problem, demanding greater efficiency, particularly in the reuse of abandoned areas, which has become an alternative. An interesting alternative would be installing energy facilities like solar, wind, biomass, and geothermal, in these areas. The objective of this paper is to develop a classification methodology, based on Artificial Intelligence (AI) and Quantum Theory (QT), to automatically carry out the classification of abandoned areas suitable for the settlement of these power plants. Artificial Neural Networks (ANNs) improved by the hybrid algorithm Quantum-behaved Particle Swarm Optimization (QPSO) together with the Levenberg-Marquardt Algorithm (LMA) were used for the classification task. In terms of Mean Squared Error (MSE), the QPSO-LMA approach achieved a decrease of 19.6% in relation to the classical LMA training with random initial weights. Moreover, the model’s accuracy showed an increase of 7.3% for the QPSO-LMA over the LMA. To validate this new approach, it was also tested on six different datasets available in the UCI Machine Learning Repository and seven classical techniques established in the literature. For the problem of installing photovoltaic plants in abandoned areas, the knowledge acquired with the solar dataset can be extrapolated to other regions.

Application of the ARIMA Model to Predict Under-Reporting of New Cases of Hansen’s Disease during the COVID-19 Pandemic in a Municipality of the Amazon Region

David Gabriel de Barros Franco — Sun, 20 Mar 2022 16:42:04 +0100

This work aimed to apply the ARIMA model to predict the under-reporting of new Hansen’s disease cases during the COVID-19 pandemic in Palmas, Tocantins, Brazil. This is an ecological time series study of Hansen’s disease indicators in the city of Palmas between 2001 and 2020 using the autoregressive integrated moving averages method. Data from the Notifiable Injuries Information System and population estimates from the Brazilian Institute of Geography and Statistics were collected. A total of 7035 new reported cases of Hansen’s disease were analyzed. The ARIMA model (4,0,3) presented the lowest values for the two tested information criteria and was the one that best fit the data, as AIC = 431.30 and BIC = 462.28, using a statistical significance level of 0.05 and showing the differences between the predicted values and those recorded in the notifications, indicating a large number of under-reporting of Hansen’s disease new cases during the period from April to December 2020. The ARIMA model reported that 177% of new cases of Hansen’s disease were not reported in Palmas during the period of the COVID-19 pandemic in 2020. This study shows the need for the municipal control program to undertake immediate actions in terms of actively searching for cases and reducing their hidden prevalence.