Jorge L. M. Amaral

Machine learning algorithms and forced oscillation measurements applied to the automatic identification of chronic obstructive pulmonary disease

The purpose of this study is to develop a clinical decision support system based on machine learning (ML) algorithms to help the diagnostic of chronic obstructive pulmonary disease (COPD) using forced oscillation (FO) measurements. To this end, the performances of classification algorithms based on Linear Bayes Normal Classifier, K nearest neighbor (KNN), decision trees, artificial neural networks (ANN) and support vector machines (SVM) were compared in order to the search for the best classifier. Four feature selection methods were also used in order to identify a reduced set of the most relevant parameters. The available dataset consists of 7 possible input features (FO parameters) of 150 measurements made in 50 volunteers (COPD, n = 25; healthy, n = 25). The performance of the classifiers and reduced data sets were evaluated by the determination of sensitivity (Se), specificity (Sp) and area under the ROC curve (AUC). Among the studied classifiers, KNN, SVM and ANN classifiers were the most adequate, reaching values that allow a very accurate clinical diagnosis (Se > 87%, Sp > 94%, and AUC > 0.95). The use of the analysis of correlation as a ranking index of the FOT parameters, allowed us to simplify the analysis of the FOT parameters, while still maintaining a high degree of accuracy. In conclusion, the results of this study indicate that the proposed classifiers may contribute to easy the diagnostic of COPD by using forced oscillation measurements.

Machine learning algorithms and forced oscillation measurements to categorise the airway obstruction severity in chronic obstructive pulmonary disease

The purpose of this study was to develop automatic classifiers to simplify the clinical use and increase the accuracy of the forced oscillation technique (FOT) in the categorisation of airway obstruction level in patients with chronic obstructive pulmonary disease (COPD). The data consisted of FOT parameters obtained from 168 volunteers (42 healthy and 126 COPD subjects with four different levels of obstruction). The first part of this study showed that FOT parameters do not provide adequate accuracy in identifying COPD subjects in the first levels of obstruction, as well as in discriminating between close levels of obstruction. In the second part of this study, different supervised machine learning (ML) techniques were investigated, including k-nearest neighbour (KNN), random forest (RF) and support vector machines with linear (SVML) and radial basis function kernels (SVMR). These algorithms were applied only in situations where high categorisation accuracy [area under the Receiver Operating Characteristic curve (AUC)≥0.9] was not achieved with the FOT parameter alone. It was observed that KNN and RF classifiers improved categorisation accuracy. Notably, in four of the six cases studied, an AUC≥0.9 was achieved. Even in situations where an AUC≥0.9 was not achieved, there was a significant improvement in categorisation performance (AUC≥0.83). In conclusion, machine learning classifiers can help in the categorisation of COPD airway obstruction. They can assist clinicians in tracking disease progression, evaluating the risk of future disease exacerbations and guiding therapy.

An immune inspired fault diagnosis system for analog circuits using wavelet signatures

This work focuses on fault diagnosis of electronic analog circuits. A fault diagnosis system for analog circuits based on wavelet decomposition and artificial immune systems is proposed. It is capable of detecting and identifying faulty components in analog circuits by analyzing its impulse response. The use of wavelet decomposition for preprocessing of the impulse response drastically reduces the size of the detector used by the Real-valued Negative Selection Algorithm (RNSA). Results have demonstrated that the proposed system is able to detect and identify faults in a Sallen-Key bandpass filter circuit.

An Immune Fault Detection System for Analog Circuits with Automatic Detector Generation

This work focuses on fault detection of electronic analog circuits. A fault detection system for analog circuits based on cross-correlation and artificial immune system is proposed. It is capable of detecting faulty components in analog circuits by analyzing its impulse response. The use of cross-correlation for preprocessing the impulse response drastically reduces the size of the detector used by the real-valued negative selection algorithm (RNSA). The proposed method can automatically generate very efficient detectors by using quadtree decomposition. Results have demonstrated that the proposed system is able to detect faults in a Sallen-Key bandpass filter and in a continuous-time state variable filter.

Towards evolvable analog artificial neural networks controllers

This work deals with the design of analog circuits for artificial neural networks (ANNs) controllers using an evolvable hardware (EHW) platform. ANNs are massively parallel systems that rely on simple processors and dense arrangements of interconnections. These networks have demonstrated their ability to deliver simple and powerful solutions in several areas, including control systems. The EHW analog platform is a reconfigurable platform, called programmable analog multiplexer array-next generation (PAMA-NG), which can be programmed by genetic algorithms to synthesize circuits. This article focuses on the development of artificial neuron circuits for analog ANNs on the PAMA-NG.

An improved method of early diagnosis of smoking-induced respiratory changes using machine learning algorithms

The purpose of this study was to develop an automatic classifier to increase the accuracy of the forced oscillation technique (FOT) for diagnosing early respiratory abnormalities in smoking patients. The data consisted of FOT parameters obtained from 56 volunteers, 28 healthy and 28 smokers with low tobacco consumption. Many supervised learning techniques were investigated, including logistic linear classifiers, k nearest neighbor (KNN), neural networks and support vector machines (SVM). To evaluate performance, the ROC curve of the most accurate parameter was established as baseline. To determine the best input features and classifier parameters, we used genetic algorithms and a 10-fold cross-validation using the average area under the ROC curve (AUC). In the first experiment, the original FOT parameters were used as input. We observed a significant improvement in accuracy (KNN=0.89 and SVM=0.87) compared with the baseline (0.77). The second experiment performed a feature selection on the original FOT parameters. This selection did not cause any significant improvement in accuracy, but it was useful in identifying more adequate FOT parameters. In the third experiment, we performed a feature selection on the cross products of the FOT parameters. This selection resulted in a further increase in AUC (KNN=SVM=0.91), which allows for high diagnostic accuracy. In conclusion, machine learning classifiers can help identify early smoking-induced respiratory alterations. The use of FOT cross products and the search for the best features and classifier parameters can markedly improve the performance of machine learning classifiers.

Automatic identification of Chronic Obstructive Pulmonary Disease Based on forced oscillation measurements and artificial neural networks

The purpose of this study is to develop an automatic classifier based on Artificial Neural Networks (ANNs) to help the diagnostic of Chronic Obstructive Pulmonary Disease (COPD) using forced oscillation measurements (FOT). The classifier inputs are the parameters provided by the FOT and the output is the indication if the parameters indicate COPD or not. The available dataset consists of 7 possible input features (FOT parameters) of 90 measurements made in 30 volunteers. Two feature selection methods (the analysis of the linear correlation and forward search) were used in order to identify a reduced set of the most relevant parameters. Two different training strategies for the ANNs were used and the performance of resulting networks were evaluated by the determination of accuracy, sensitivity (Se), specificity (Sp) and AUC. The ANN classifiers presented high accuracy (Se > 0.9, Se > 0.9 and AUC > 0.9) both in the complete and the reduce sets of FOT parameters. This indicates that ANNs classifiers may contribute to easy the diagnostic of COPD using forced oscillation measurements.

Real-valued negative selection algorithm with a Quasi-Monte Carlo genetic detector generation

A new scheme for detector generation for the Real-Valued Negative Selection Algorithm (RNSA) is presented. The proposed method makes use of genetic algorithms and Quasi-Monte Carlo Integration to automatically generate a small number of very efficient detectors. Results have demonstrated that a fault detection system with detectors generated by the proposed scheme is able to detect faults in analog circuits and in a ball bearing dataset.

Evolvable building blocks for analog fuzzy logic controllers

This work discusses the use of an evolvable hardware (EHW) platform in the synthesis of analog electronic circuits for fuzzy logic controllers. A fuzzy logic controller (FLC) is defined by a collection of fuzzy if-then rules and a set of membership functions characterizing the linguistic terms associated with the inputs and output of the FLC. The EHW analog platform, named PAMA-NG (programmable analog multiplexer array - next generation), is a reconfigurable platform that consists of integrated circuits whose internal connections can be programmed by evolutionary computation techniques, such as genetic algorithms, to synthesize circuits. The PAMA-NG is classified as a field programmable analog array (FPAA). FPAAs have appeared recently and constitute the state of the art in the technology of reconfigurable platforms. These devices will become the building blocks of a forthcoming class of hardware, with the important features of self-adaptation and self-repairing, through automatic reconfiguration. This article focuses on the development of building blocks for analog FLCs on the PAMA-NG and presents case studies.

Fault detection in analog circuits using a fuzzy dendritic cell algorithm

This work presents the early stages of the development of a fault detection system based on the Dendritic Cell Algorithm. The system is designed to detect parametric faults in linear time invariant circuits. The safe signal is related to the mean square error between the PAA representations of the impulse responses of the circuit under test and the golden circuit. The danger signal is related to the variation of that error. Instead of using a weighted sum with fixed weights, a fuzzy inference system (FIS) is used, since it is easier to define linguistic rules to infer the combination of the signals than to find appropriate weight values.