Fisher Discriminant Ratio Research Articles

Optimized data driven fault detection and diagnosis in chemical processes

Fault detection and diagnosis (FDD) is crucial for ensuring process safety and product quality in the chemical industry. Despite the large amounts of process data recorded and stored in chemical plants, most of them are not well-labeled, and their conditions are not adequately specified. In this study, an optimized data-driven FDD model was developed for a chemical process based on automatic clustering algorithms. Due to data preprocessing importance, feature selection was performed by a non-dominated sorting genetic algorithm (NSGAII) based on k-means clustering. The optimal subset of features is selected by comparing clustering results for each subset. The performance of the proposed feature selection method was compared with the Fisher discriminant ratio (FDR), and XGBoost methods. The t-distributed stochastic neighbor embedding (t-SNE), Isomap, and KPCA dimension reduction methods were also employed for feature extraction. Finally, automatic clustering was performed based on metaheuristic algorithms for fault detection and diagnosis. Results were compared with non-automatic clustering methods. The performance of the proposed method was evaluated by examining the Tennessee Eastman and four water tank processes as case studies. The results showed that the proposed method is reliable and capable of online and offline chemical process fault detection and diagnosis. As a result, the findings of this study can be used to stabilize the operation of chemical processes.

Segmentability evaluation of back-scattered SEM images of multiphase materials

Segmentation methods are very useful tools in the Electron Microscopy inspection of materials, enabling the extraction of quantitative results from microscopy images. Back-Scattered Electron (BSE) images carry information of the mean atomic number in the interaction volume and hence can be used to quantify the phase composition in multiphase materials. Since phase composition and proportion affects the material properties and hence its applications, the segmentation accuracy of such images rendered of critical importance for material science. In this work, the notion of segmentability for BSE images is proposed to define the ability of an image to be segmented accurately. This notion can be used to guide the image acquisition process so that segmentability is maximized and segmentation accuracy is ensured. An index is devised to quantify segmentability based on a combination of the modified Fisher Discrimination Ratio and of the second Minkowski functional capturing intensity and spatial aspects of BSE images respectively. The suggested Segmentability Index (SI) is validated in synthetic BSE images which are generated with a novel algorithm allowing the independent control of spatial distribution of phases and their grayscale intensity histograms. Additionally, SI is applied in real-synthetic BSE images, where the real greyscale distributions of Ordinary Portland Cement (OPC) clinker crystallographic phases are used, to demonstrate the ability of SI to indicate the optimum choice of critical image acquisition settings leading to the more accurate segmentation output.

CARDIAC DISEASE PREDICTION USING RANDOM FOREST WITH LINEAR MODEL

Making forecasts and diagnosing ailments has never been simple for medical professionals when it comes to heart conditions. Due to this, people can take the necessary action to treat heart disease before it gets worse if it is discovered in its early stages anywhere in the world. The three main causes of heart disease-drinking alcohol, smoking cigarettes, and not exercising-have become serious issues in recent years. The health care industry has produced a substantial amount of data over time, which has made machine learning capable of providing effective outcomes in prediction and decision-making. Only male patients risk level for heart disease is predicted by the Heart Disease Prediction (HDP) in the current system, which is created using the Naive Bayes and Decision Tree algorithms. For prediction, the algorithm makes use of medical parameters such age, sex, smoking status, BMI, and physical health, etc. A patients likelihood of developing heart disease is predicted by the HDP. Fishers Discriminant Ratio is one of the methods for feature selection based on random forest (RF), which is recommended to identify the best features for heart disease prediction and enhance the precision of RF-based classification. Our goal in this study is to identify the best factors that can improve the prediction accuracy of heart disease and finding the most effective variables to raise the accuracy of heart disease prediction. Evaluation criteria, namely accuracy, specificity, sensitivity, and area under the ROC curve, are employed to verify the efficacy of the proposed approach on a public dataset comprising patients of both genders. The primary benefits of applying machine learning for heart disease prediction are that it reduces the complexity of the doctors time, is patient- and cost-friendly, and manages the largest (enormous) amount of data through feature selection and the random forest algorithm. Early diagnosis of cardiovascular disease can help with lifestyle modifications for high-risk patients, which can lower complications and be a significant medical milestone. KEYWORDS: Random Forest, Decision Tree, Naive Bayes, Feature Selection, and Evaluation Metrics.

Analytical chemistry kernel library for spectroscopic profiling data

The kernel is a powerful mathematical tool in spectroscopic profiling data analysis. This paper gives a theoretical and systematic formulation for various kernel types. An open-source Python toolkit (ackl, “analytical chemistry kernel library”) is developed accordingly. Highlights of the toolkit include: (1) It designs a unified API and implements totally 31 kernel types (e.g., linear, poly, Gaussian, Matern, Cauchy, Sorensen, wavelet, Fejér, etc.), which is by far the most comprehensive kernel library. (2) It provides a tailored hyper-parameter optimization mechanism for each kernel, which suits the spectroscopic profiling data properties. (3) It designs a set of 16 evaluation metrics (e.g., classification accuracy, F1 score, MANOVA test statistic, Kolmogorov-Smirnov test statistic, Cohen effect size, Fisher's discriminant ratio, computational cost, etc.) to compare different kernels in discriminative tasks. Finally, we conducted spectroscopic profiling case studies using this tool and summarized a general guideline for kernel selection.

BoF-SVM-based data intelligence model for detecting dementia

Dementia is a brain condition that impairs the cognitive abilities of an individual. Mild cognitive impairment is a mediator phase of healthy and dementia controls. The motivation of this study is to predict dementia using magnetic resonance imaging data, which is significant for the diagnosis of normal control and dementia patients. The proposed model leverages effective methods like Discrete Wavelet Transform, Bag of Features, and Support Vector Machine. The four wavelets haar, Daubechies, symlets, and coiflets are used for image compression. The results of the proposed data intelligence model are promising in terms of accuracy which is 92.32% which is better than the recently proposed models. Also, the proposed data intelligence model is compared with the models which may use curvelet transform, and shearlet transform and with the methods which have gone without using DWT transforms. The comparisons have also been made with the models that have used other prevalent techniques like Principal Component Analysis, Fisher Discriminant Ratio, and Gray Level Co-occurrence Matrix. The outcomes support the usage of each technique in the suggested data intelligence paradigm.

An algorithmic trading system based on a stacked generalization model and hidden Markov model in the foreign exchange market

The Forex market has been one of the most attractive markets to researchers, funds, and traders. Literature shows that a single model algorithm usually cannot perform satisfactorily on the foreign exchange (Forex) time series because of the market's complexity. This study develops an algorithm based on two stacked generalization models consisting of four machine-learning models. First, the optimal lags of features are found using the Fisher discriminant ratio and partial autocorrelation function. Second, one stacked model fits the bullish trends, and the other holds the bearish trends resulting from a hidden Markov model. We reinforce the predictive signals of these models by extracting relationships between currency pairs with correlation and mutual information. Lastly, the proposed algorithm constructs a portfolio based on the strength of signals dependent on correlation and mutual information. As a result, this paper compares the performance of the proposed approach with different states of the model and several established benchmarks regarding return and risk metrics. The outcomes show that the proposed model's added features—such as time series clustering, considering a range of inputs, and internal relationships among different assets—can increase its performance in the Forex market.

Guided waves based debonding classification in lap-joints using modified Fisher discriminant criterion

In this paper, a modified fisher discriminant criterion is used to identify and rank damage sensitive features for debonding classification. For this purpose, the analysis of the correlation between features (ACF) based on Pearson correlation coefficient and the Fisher discriminant ratio (FDR) criterion were combined to select the most relevant features. The debondings classification was evaluated in a multiclass classification by considering three different classes of damage and one class of undamaged. The nondestructive inspection was performed in a composite-adhesive-composite lap-joint using guided Lamb wave testing with angle beam transducers mounted in a pitch-catch configuration. From the collected data, twelve features were extracted by using statistical functions and signal processing techniques. From the extracted features, the most damage sensitive features were selected based on the proposed method (ACF and FDR) and the results were compared using the conventional FDR method. For three different classifiers, the feature selection strategies were tested and their performance evaluated using the metrics: accuracy, precision, recall and F1 score. For the dataset collected in this study, the results support that the proposed method yields an improvement of the performance of the classifiers in comparison with the use of the conventional FDR-based feature selection.

A Statistical Signal Processing Approach to Islanding Detection

The integration of distributed generation (DG) sources in the electric energy systems may bring new problems that need attention, one of these problems is the occurrence of unintentional islanding. Islanding is a condition in which part of the distribution network is disconnected from the system, and consumer units are still powered by one or more DGs, which can cause damage to equipment and pose risks to the safety of technicians. This paper shows an islanding detection method (IDM) in Power Systems with DG based on statistical signal processing. We used a MathWorks Simulink model of a grid-connected 250 kW photovoltaic (PV) array to simulate the behavior of the three-phase voltage signal in the point of common coupling (PCC) under the nominal operation, islanding condition, and fault condition using different load compositions. Principal Component Analysis (PCA) was used to extract the transitory events from the voltage signals, and then we used second-, third-, and fourth-order cumulants to generate features and the best ones were selected using the Fisher’s Discriminant Ratio (FDR). A Radial Basis Function Network (RBFN) makes the classification of the events. We found that, for this setup, we can achieve detection rates of 99% for both islanding condition detection and fault occurrence classification, no matter the power mismatch between the load and the DG.

Heart sound classification based on improved mel frequency cepstrum coefficient and integrated decision network method

Heart sound analysis is significant for early diagnosis of congenital heart disease. A novel method of heart sound classification was proposed in this paper, in which the traditional mel frequency cepstral coefficient (MFCC) method was improved by using the Fisher discriminant half raised-sine function (F-HRSF) and an integrated decision network was used as classifier. It does not rely on segmentation of the cardiac cycle. Firstly, the heart sound signals were framed and windowed. Then, the features of heart sounds were extracted by using improved MFCC, in which the F-HRSF was used to weight sub-band components of MFCC according to the Fisher discriminant ratio of each sub-band component and the raised half sine function. Three classification networks, convolutional neural network (CNN), long and short-term memory network (LSTM), and gated recurrent unit (GRU) were combined as integrated decision network. Finally, the two-category classification results were obtained through the majority voting algorithm. An accuracy of 92.15%, sensitivity of 91.43%, specificity of 92.83%, corrected accuracy of 92.01%, and F score of 92.13% were achieved using the novel signal processing techniques. It shows that the algorithm has great potential in early diagnosis of congenital heart disease.

Reducing Data Complexity Using Autoencoders With Class-Informed Loss Functions.

Available data in machine learning applications is becoming increasingly complex, due to higher dimensionality and difficult classes. There exists a wide variety of approaches to measuring complexity of labeled data, according to class overlap, separability or boundary shapes, as well as group morphology. Many techniques can transform the data in order to find better features, but few focus on specifically reducing data complexity. Most data transformation methods mainly treat the dimensionality aspect, leaving aside the available information within class labels which can be useful when classes are somehow complex. This paper proposes an autoencoder-based approach to complexity reduction, using class labels in order to inform the loss function about the adequacy of the generated variables. This leads to three different new feature learners, Scorer, Skaler and Slicer. They are based on Fisher's discriminant ratio, the Kullback-Leibler divergence and least-squares support vector machines, respectively. They can be applied as a preprocessing stage for a binary classification problem. A thorough experimentation across a collection of 27 datasets and a range of complexity and classification metrics shows that class-informed autoencoders perform better than 4 other popular unsupervised feature extraction techniques, especially when the final objective is using the data for a classification task.

A Data-Driven Approach for Assessing Aero-Engine Health Status

Aero-engine health assessment is an important issue in the prognostics and health management (PHM) field. To improve the accuracy of health assessment, a novel data-driven approach is proposed in this paper. Unlike the model-based methods, the proposed data-driven method does not require to know the physical nature of degradation mechanism. Also, it is different from traditional data-driven methods and does not require to construct the synthesized health index or pre-set the health thresholds. In this work, the proposed method consists of feature selection, health state division and health state evaluation. Firstly, Fisher's discriminant ratio (FDR) is proposed to evaluate the sensor-based condition monitoring information, so as to eliminate the redundant information while retaining dominant information. Secondly, to solve the problem of unlabeled observation data, fuzzy c-means (FCM) clustering algorithm is presented to define the aero-engine health states. Finally, a bidirectional long short term memory (Bi-LSTM) network, which can capture the bidirectional long-range dependencies of features, is introduced to construct the health state evaluation model. Experimental results on aero-engine data set from NASA indicate that the proposed method can effectively improve the accuracy of health assessment, which can inform maintenance-related decisions.

High-impedance fault modeling and classification in power distribution networks

This work presents a novel method for the classification of high-impedance faults (HIFs) in power distribution networks based on the association of higher-order statistics (HOS) and a multilayer perceptron (MLP) artificial neural network (ANN). An alternative model is developed to represent the HIF phenomenon considering five different contact surfaces with the ground. A broad analysis comprising six types of typical events that occur in distribution networks is performed in the Alternative Transient Program (ATP) software, including several conditions such as: normal operation; single-phase faults; two-phase faults; three-phase faults; energization of transformers and capacitor banks; switching of inductive loads; as well as faults involving five modeled surfaces. HOS is combined with Fisher’s discriminant ratio (FDR) to extract the best characteristics. At the end, the MLP-type ANN is used to recognize the specific patterns of each event aiming to identify each event accurately, especially HIFs. The obtained results demonstrate that the proposed technique proves to be a reliable and accurate tool, achieving classification hits above 98%.

Machine Learning Classifiers for Surface Crack Detection in Fracture Experiments

Correctly determining the onset of fracture is crucial when performing mechanical experiments. Commonly carried out by visual inspection, here an image-based machine learning approach is proposed to classify cracked and un-cracked specimens. It yields the potential to objectify and automate crack detection, thereby removing sources of uncertainty and error from the post-processing of experiments. More than 30’000 speckle-pattern images obtained from 77 experiments on three specimen geometries are evaluated. They comprise uniaxial tension, notched tension as well as axisymmetric V-bending experiments. Statistical texture features are extracted from all images. They include both first-order (variance, skewness, kurtosis) and higher-order statistical texture features, i.e. Haralick features. The discriminatory power of the texture information is evaluated based on the Fisher's Discriminant Ratio and feature correlations are identified and quantified. Image texture feature subsets of high discriminatory power are used to parse neural network architectures of different complexities from simple perceptron to feed-forward and cascade neural networks. It is found that a small subset of the investigated texture features is highly significant for all experiments. Using this feature subset in conjunction with multi-layer, non-linear and low complexity feed-forward network architectures classification accuracies in the order of 99% are obtained. At the same time, it is shown that linear classifiers are not sufficient to robustly distinguish the state of the specimens, even when high discriminatory power features are used.Graphical Abstract: [Display omitted]

Emotion Recognition Using Facial Expressions

Emotion recognition is an important aspect of human interaction, and this ability of humans to interpret emotions based on facial expressions is a basic element for effective communication. Machine learning can help automate this complicated task with the help of feature engineering. This work proposes some pipelines trained on the JAFFE dataset using feature extraction methods, namely principal component analysis (PCA) and local binary pattern (LBP) combined with Fisher discriminant ratio (FDR) as a feature selection method. In order to build a classification scheme capable of successfully identifying face images related to the six universal emotions and neutral expression, all possible combinations have been empirically analyzed. In the final model, PCA combined with FDR has been used on the support vector machine classifier with a linear kernel. The results obtained are encouraging and this work may also prove important for disciplines other than computer science such as for management purposes.

On effective cognitive state classification using novel feature extraction strategies.

Investigating new features for human cognitive state classification is an intiguing area of research with Electroencephalography (EEG) based signal analysis. We plan to develop a cost-effective system for cognitive state classification using ambulatory EEG signals. A novel event driven environment is created using external stimuli for capturing EEG data using a 14-channel Emotiv neuro-headset. A new feature extraction method, Gammatone Cepstrum Coefficients (GTCC) is introduced for ambulatory EEG signal analysis. The efficacy of this technique is compared with other feature extraction methods such as Discrete Wavelet Transformation (DWT) and Mel-Frequency Cepstral Coefficients (MFCC) using statistical metrics such as Fisher Discriminant Ratio (FDR) and Logistic Regression (LR). We obtain higher values for GTCC features, demonstrating its discriminative power during classification. A superior performance is achieved for the EEG dataset with a novel ensemble feature space comprising of GTCC and MFCC. Furthermore, the ensemble feature sets are passed through a proposed 1D Convolution Neural Networks (CNN) model to extract novel features. Various classification models like Probabilistic neural network (P-NN), Linear Discriminant Analysis (LDA), Multi-Class Support Vector Machine (MCSVM), Decision Tree (DT), Random Forest (RF) and Deep Convolutional Generative Adversarial Network (DCGAN) are employed to observe best accuracy on extracted features. The proposed GTCC, (GTCC+MFCC) & (GTCC +MFCC +CNN) features outperform the state-of-the-art techniques for all cases in our work. With GTCC+MFCC feature space and GTCC+MFCC+CNN features, accuracies of 96.42% and 96.14% are attained with the DCGAN classifier. Higher classification accuracies of the proposed system makes it a cynosure in the field of cognitive science.

Classification of Toddler Nutritional Status Based on Antrophometric Index and Feature Discrimination using Support Vector Machine Hyperparameter Tuning

Nutritional status is the study of food and is related to health. Nutritional status is a benchmark to assess the health development of toddler. The nutritional status of toddler is assessed according to three index, such as body weight to age (BW / A), body height to age (BH / A), body weight to body height (BW / BH). The issue of nutrition is still a major factor in the growth and development of toddler in Indonesia. Public Health Center (Puskesmas) and Integrated Healthcare Center (Posyandu) as public health services work together to control the growth and development of toddler in Indonesia. To help control the growth and development of toddler, we proposed a research to classify the nutritional status of toddler based on anthropometric index. The nutritional status of toddler dataset was formed into a classification model using SVM Hyperparameter Tuning. SVM is a machine learning which the classification model used a hypothesis space in the form of linear functions in a high dimensional feature space. Adjustment of the hyperparameter was involved to reach a model that can optimally solve machine learning problems. We implemented feature selection using Fisher's Discriminant Ratio as a preprocessing stage, which the most important features were body weight (BB) and height (BH). The experimental results showed the classification model using SVM on training and testing data with a ratio of 70:30 reached accuracy of 84%, while SVM Hyperparameter Tuning with parameter of Cost = 100 parameters, Gamma = 0.01, Kernel = RBF reached accuracy of 97%. They represented a significant accuracy difference of 13%.

Writer identification using redundant writing patterns and dual-factor analysis of variance

Writer identification (WI) is a typical pattern recognition problem with the goal of recognizing the writer of a text from images of his or her handwriting. For handwriting-based applications, a new approach is required that can confirm the writer based on a very small amount of available text. However, due to the limited number of characters and the diverse contents of the available samples, it is still challenging to achieve a high recognition rate on large-scale databases. To solve this problem, this paper proposes an efficient method for offline text-independent WI using redundant writing patterns and dual-factor analysis of variance (DF-ANOVA). In this method, a search for writing patterns in a limited number of handwritten texts is first conducted to generate meaningful codebooks. Then, these writing patterns are described by means of the simplified Wigner distribution function (WDF) and improved directional index histogram (DIH) descriptors to extract both structural and texture features, and the two corresponding feature distances are integrated by means of a fuzzy integral rule to make the final decision. The Fisher discriminant ratio (FDR) is used for feature selection, and DF-ANOVA is used to eliminate the influence of factors associated with the labels of samples on the feature distance. The results of evaluations on the IAM (96.92%), Firemaker (96.4%), and Uyghur2016 (100%) databases illustrate that this system shows strong robustness to an increasing number of writers and a decreasing number of words in the sample texts.

A Study of EEG Feature Complexity in Epileptic Seizure Prediction

The purpose of this study is (1) to provide EEG feature complexity analysis in seizure prediction by inter-ictal and pre-ital data classification and, (2) to assess the between-subject variability of the considered features. In the past several decades, there has been a sustained interest in predicting epilepsy seizure using EEG data. Most methods classify features extracted from EEG, which they assume are characteristic of the presence of an epilepsy episode, for instance, by distinguishing a pre-ictal interval of data (which is in a given window just before the onset of a seizure) from inter-ictal (which is in preceding windows following the seizure). To evaluate the difficulty of this classification problem independently of the classification model, we investigate the complexity of an exhaustive list of 88 features using various complexity metrics, i.e., the Fisher discriminant ratio, the volume of overlap, and the individual feature efficiency. Complexity measurements on real and synthetic data testbeds reveal that that seizure prediction by pre-ictal/inter-ictal feature distinction is a problem of significant complexity. It shows that several features are clearly useful, without decidedly identifying an optimal set.

Investigation and Evaluation of Glottal Flow Waveform for Voice Pathology Detection

Automatic voice pathology detection can provide objective estimation and prevention in the early stages of voice diseases. Glottal flow waveform directly reflects the state of glottal excitation. Extracting acoustic features from glottal source signals may contribute to the detection of pathological voice. To improve the performance of voice pathology detection, this article investigates the contribution of the glottal flow waveform for pathological voice detection by evaluating the classification result using features extracted from raw speech utterances and corresponding glottal flow waveforms. The individual feature sets used are extracted from raw or glottal voice utterances with identical parameter settings, which are openSMILE acoustic features, audio features computed by Moving Picture Experts Group-7 standard and classical glottal source features. In addition, a feature selection method in terms of the wrapper approach is used to combine the single features ranked by using the Fisher discrimination ratio. Voice pathology detection experiments were carried out using Random Forest. The best accuracies of 88.52% for the Saarbrucken Voice database and 100.00% for the Massachusetts Eye and Ear Infirmary database are achieved using the combined feature set extracted from the glottal source signal, with improvement of 0.44-3.13% in the accuracies obtained by using raw speech utterances. Compared to state-of-the-art methods, the proposed method achieves the highest accuracy for the Massachusetts Eye and Ear Infirmary database and an increase of 2.75-17.16% in detection accuracy compared to other conventional pipeline systems for the Saarbrucken Voice databse. The experimental results demonstrate that using glottal flow waveform as source signal can improve the performance of pathological voice detection.

A Feature-Driven Decision Support System for Heart Disease Prediction Based on Fisher's Discriminant Ratio and Backpropagation Algorithm

Coronary heart disease included a group of cardiovascular, and it is a leading cause of death in low and middle-income countries. Risk factors for coronary heart disease are divided into two, namely primary and secondary risk factors. The need to identify characteristics or risk factors in heart disease patients by making the classification model. The modeling of heart disease classification to know how the system can able to reach the best prediction accuracy. Fisher's Discriminant Ratio is one of the methods for feature selection, which is used to get high discriminant features. While Backpropagation is one of the classification models to recognize patterns in heart disease patients. The experiment results showed that the accuracy of the classification model using 13 original features reached 92%. By reducing the features based on the score of the feature selection, then the lowest feature was removed from original features and left there were 12 features involved in the classification model which the accuracy increased to 93%. Furthermore, the results of determining the threshold (accuracy does not decrease continuously) and consider the effect of eliminating the lowest features that are considered quite fluctuating on accuracy. The accuracy reached 90% by eliminating the five lowest features and left eight existing features.