Parkinson Disease Identification Using Residual Networks and Optimum-Path Forest

EEG signal classification for epilepsy diagnosis via optimum path forest – A systematic assessment

Machine learning techniques have been extensively researched in the last years, mainly due to their effectiveness when dealing with recognition or classification applications. Typically, one can comprehend using a Machine Learning system to autonomously delegate routines, save human efforts, and produce great insights regarding decision-making tasks. This paper introduces and validates a stacking-based ensemble approach using Optimum-Path Forest classifiers in intrusion detection tasks. Instead of only using the famous NSL-KDD dataset, we propose a new dataset called uneSPY, which we believe will fill the gap concerning new intrusion detection datasets. Both datasets were evaluated under several classifiers, including Logistic Regression, Decision Trees, Support Vector Machines, Optimum-Path Forests, and compared against Optimum-Path Forest stacking-based ensembles. Experimental results showed an Optimum-Path Forest stacking-based ensemble classification suitability, particularly when considering its ability to generalize large volumes of data while sustaining its performance.

In this paper, the Optimum-Path Forest (OPF) classifier is applied in the classification of defects in cowhide, a problem of great evaluation complexity. The OPF classifier reduces a pattern classification problem to the problem of partitioning the vertices of a graph induced by its data set. The results revealed a competent performance compared to traditional classifiers, such as Support Vector Machines (SVM), Artificial Neural Networks-Perceptron Multilayer (MLP), Decision Trees (J48) and k-Nearest Neighbor (kNN).

The medical diagnosis of Parkinson’s disease (PD) mainly relies on doctors’ clinical experience and scoring scales, which may lead to erroneous judgments due to the differences of their clinical experience. Gait information, such as postural instability and bradykinesia, is a common symptom in PD. Gait analysis and classification is often adopted as a clinical tool for detecting and diagnosing PD. In this paper, gait parameters are collected from 200 PD patients and 100 healthy controls (HC) through the wearable sensors, which include 95 features. Considering the redundancy between the features, F -test and Recursive Feature Elimination (RFE) methods are utilized for feature selection. Moreover, support vector machine (SVM) algorithm is utilized to train the classifier, which can automatically classify PD patients and HC. Furthermore, experiments on the class-imbalance and class-equilibrium samples are performed to verify the reliability and stability of the above methods. Compared with all feature based SVM and F -test based SVM, the RFE-based SVM has better performance, where the accuracy, sensitivity and specificity are 96.67%, 96.77% and 96.55%, respectively. The results show that the RFE-based SVM classifier can well distinguish PD and healthy gait, and can effectively assist doctors to diagnose PD in clinical practice, thus breaking the subjectivity brought by traditional PD diagnosis.

Cutting analysis is a important and crucial task task to detect and prevent problems during the petroleum well drilling process. Several studies have been developed for drilling inspection, but none of them takes care about analysing the generated cutting at the vibrating shale shakers. Here we proposed a system to analyse the cutting's concentration at the vibrating shale shakers, which can indicate problems during the petroleum well drilling process, such that the collapse of the well borehole walls. Cutting's images are acquired and sent to the data analysis module, which has as the main goal to extract features and to classify frames according to one of three previously classes of cutting's volume. A collection of supervised classifiers were applied in order to allow comparisons about their accuracy and efficiency. We used the Optimum-Path Forest (OPF), Artificial Neural Network using Multilayer Perceptrons (ANN-MLP), Support Vector Machines (SVM) and a Bayesian Classifier (BC) for this task. The first one outperformed all the remaining classifiers. Recall that we are also the first to introduce the OPF classifier in this field of knowledge. Very good results show the robustness of the proposed system, which can be also integrated with other commonly system (Mud-Logging) in order to improve the last one's efficiency.

Some pattern recognition techniques may present a high computational cost for learning samples' behaviour. The Optimum-Path Forest (OPF) classifier has been recently developed in order to overcome such drawbacks. Although it can achieve faster training steps when compared to some state-of-art techniques, OPF can be slower for testing in some situations. Therefore, we propose in this paper an implementation in graphics cards of the OPF classification, which showed to be more efficient than traditional OPF with similar accuracies.

Parkinson's disease (PD) is a neurological disorder which is progressive in nature. Although there is no cure to this disease, symptomatic treatments are available. These treatments can slow the progressive development of the symptoms. Medications can treat some of the symptoms of the PD up to a great extent that in turn may help the patients to live a normal life. Besides these medications, the patients can also be provided with various therapies based on the types of their symptoms. But for providing any treatment, detection of its symptoms at an early stage is very important. This can reduce its future complexities considerably. Early diagnosis along with proper medications may treat the symptoms of PD significantly. This motivates to propose a new and effective methodology for detection and analysis of PD. In this work, an approach has been proposed for identification of PD patients by using Electroencephalogram (EEG) signals. Here, the EEG signals of normal persons and PD patients are processed in three stages. First, the raw EEG signals are pre-processed for removal of noises and artefacts present. Out of various techniques, Wavelet transform is used for this purpose. In the MATLAB environment, de-noising can be executed by using the in-built functions. Performances of the de-noising techniques are examined with the performance parameters namely Root Mean Square Error (RMSE) as well as Signal to Noise Ratio (SNR). In the second stage, statistical features are extracted from the pre-processed EEG signals. In this work, five statistical features are considered for performing the classification. In the final stage, the extracted features are classified using Artificial Neural Network (ANN) and Support Vector Machine (SVM) techniques. ANN is an efficient classifier that predicts the human brain's working manners. SVM on the other hand has been proven as one of the most prevailing classification algorithms that gives highly accurate and robust results. This is a novel approach of analyzing the performances of the classification techniques by evaluating the best performing feature. In both the classifiers accuracy, precision, sensitivity and specificity are calculated from the confusion matrix evaluated from the values of the statistical features. In ANN, results using six different training algorithms at different hidden layers are calculated and compared. This proves the training algorithm Levenberg-Marquardt back-propagation with hidden layer 20 as the best combination for performing the classification. From the results it is seen that both ANN and SVM classifiers provide significant classification accuracies of 94.7% and 96.5% respectively. Amongst the five considered features, Mean performs the best in terms of classification accuracy.

Could Parkinson’s disease be diagnosed at an early stage by measuring rest tremor under stressed conditions?

The use of energy sources boosts technological development over time, but the continuous and unrelenting usage of fossil fuel energy has resulted in harmful impacts to the ecosystem. Then came an alert in the scientific community, generating discourses and research in the field of renewable energy, for example, wind, solar and biofuels. Among various fields of study in wind power, this work is concerned in the study of electric generators, specifically in the detection of faults, using magnetic axial flux acquired from a squirrel cage induction generator (SCIG). We compare the performance of classifiers Bayes, Support Vector Machines, Optimum Path Forest and K- Nearest Neighbor in three different feature extraction methods, High Order Statistics (HOS), Fourier Transform and Structural Co-occurrence Matrix (SCM). We were able to correctly identify 99.4% of the Normal conditions using a Bayesian classifier alongside with as well identify incipient short-circuits with 95.6% of hit rate. SVM had a lowers false positive, 0.6% and 2% in HOS and SCM, respectively, while identifying 98% of the Normal conditions, so is more suitable for binary classification between fault and non-fault short-circuits.

When the photovoltaic (PV) system fails, the power generation efficiency is greatly reduced or even leads to incidents, that is why the fault diagnosis of the PV system is very necessary. This paper proposed an optimum-path forest (OPF) photovoltaic arrays fault diagnosis method. According to the characteristics of the photovoltaic panel, seven factors were selected as the characteristic parameters for studying the fault diagnosis of PV system. Then, combined with the actual measurement data and the simulated measurement data, the complete graphs and the minimum spanning trees were obtained in turn, and finally the OPF classifier was obtained which was used for fault diagnosis and classification. At the same time, this paper also compared the OPF algorithm with other algorithms such as Linear SVM, Fine KNN, Linear discriminant and Boosted Trees that showed the OPF algorithm has very high speed and accuracy and is more suitable for fault diagnosis of PV system in practical application. In addition, this paper also optimized the traditional OPF method which improves the speed of classification. Finally, the actual fault detection of PV system was built and achieved good fault diagnosis effect.

Machine learning (ML) and image processing techniques have been applied together to various scenarios for the development of Intelligent Vehicles. Among these scenarios, pedestrian detection has received growing interest in recent years, since high concern for safety applications in traffic has arisen. Several ML methods were successfully applied to solve this problem. However, because pedestrian detection is in general computationally intensive, a good trade off between accuracy and processing time is desirable, particularly if the methods are directed to real-time applications. Optimum Path Forest (OPF) classifier is a recently developed non-parametric classifier method. This work contribution is the performance assessment of a novel OPF application to pedestrian detection. Results have shown that it is fast and competitive against established methods and a viable alternative to be considered for machine learning and pedestrian detection applications.

Appropriate metrics are paramount for machine learning and pattern recognition. In Content-based Image Retrieval-oriented applications, low-level features and pairwise-distance metrics are usually not capable of representing similarity among the objects as observed by humans. Therefore, metric learning from available data has become crucial in such applications, but just a few related approaches take into account the contextual information inherent from the samples for a better accuracy performance. In this paper, we propose a novel approach which combines an unsupervised manifold learning algorithm with the Optimum-Path Forest (OPF) classifier to obtain more accurate recognition rates, as well as we show it can outperform standard OPF-based classifiers that are trained over the original manifold. Experiments conducted in some public datasets evidenced the validity of metric learning in the context of OPF classifiers.

Parkinson’s disease (PD) is traditionally viewed as a mainly hypodopaminergic syndrome, with symptoms resulting predominantly from loss of dopamine-producing neurons in the substantia nigra. However, while most of the cardinal motor features of PD respond well to dopaminergic therapy, many other features of the disease do not. Balance impairment and the associated risk of falling represent one of the most prominent and potentially disabling features that are typically refractory to dopaminergic treatment. Therefore, it is possible that lesions in nondopaminergic systems contribute to the pathophysiology of postural instability in PD. Such nondopaminergic lesions are well recognized, certainly in advanced stages of PD where postural instability and falls dominate the clinical presentation. However, it remains unclear which of the identified nondopaminergic lesions is specifically responsible for postural instability and balance impairment. In this review, we argue that cell loss in the locus coeruleus and a resultant central norepinephrine deficit are intimately involved in the pathophysiology of postural instability in PD. If proven to be correct, this link between defective noradrenergic neurotransmission and postural instability could have important implications for the future development of new symptomatic treatments aimed to correct postural instability and preventing falls. Studies in the next 5 years could test this hypothesis, using a battery of complementary research techniques, including advanced neuroimaging (structural, functional imaging and nuclear), neurochemical studies of cerebrospinal fluid, post-mortem clinicopathological analyses and detailed clinical balance evaluations supplemented by posturography studies.

Although one can find several pattern recognition techniques out there, there is still room for improvements and new approaches. In this book chapter, we revisited the Optimum-Path Forest (OPF) classifier, which has been evaluated over the last years in a number of applications that consider supervised, semi-supervised and unsupervised learning problems. We also presented a brief compilation of a number of previous works that employed OPF in different research fields, that range from remote sensing image classification to medical data analysis.

The research on multiple classifiers systems includes the creation of an ensemble of classifiers and the proper combination of the decisions. In order to combine the decisions given by classifiers, methods related to fixed rules and decision templates are often used. Therefore, the influence and relationship between classifier decisions are often not considered in the combination schemes. In this paper we propose a framework to combine classifiers using a decision graph under a random field model and a game strategy approach to obtain the final decision. The results of combining Optimum-Path Forest (OPF) classifiers using the proposed model are reported, obtaining good performance in experiments using simulated and real data sets. The results encourage the combination of OPF ensembles and the framework to design multiple classifier systems.