Classification of Diesel Engine Health Using Sparse Linear Discriminant Analysis (SLDA)

2009 ◽  
Author(s):  
Neha Chandrachud ◽  
Ravindra Kakade ◽  
Peter H. Meckl ◽  
Galen B. King ◽  
Kristofer Jennings
Keyword(s):  
Steady State ◽  
Diesel Engine ◽  
Discriminant Analysis ◽  
Linear Discriminant Analysis ◽  
Optimal Number ◽  
Classification Model ◽  
Misclassification Rate ◽  
Linear Discriminant ◽  
State Data ◽  
Input Variables

With requirements for on-board diagnostics on diesel engines becoming more stringent for the coming model years, diesel engine manufacturers must improve their ability to identify fault conditions that lead to increased exhaust emissions. This paper proposes a statistical classifier model to identify the state of the engine, i.e. healthy or faulty, using an optimal number of sensors based on the data acquired from the engine. The classification model proposed in this paper is based on Sparse Linear Discriminant Analysis. This technique performs Linear Discriminant Analysis with a sparseness criterion imposed such that classification, dimension reduction and feature selection are merged into one step. It was concluded that the analysis technique could produce 0% misclassification rate for the steady-state data acquired from the diesel engine using five input variables. The classifier model was also extended to transient operation of the engine. The misclassification rate in the case of transient data was reduced from 31% to 26% by using the steady-state data trained classifier using thirteen variables.

scholarly journals Linear Discriminant Analysis Successfully Predicts Knee Injury Outcome From Biomechanical Variables

2020 ◽  
Vol 48 (10) ◽  
pp. 2447-2455
Author(s):  
Nathan D. Schilaty ◽  
Nathaniel A. Bates ◽  
Sydney Kruisselbrink ◽  
Aaron J. Krych ◽  
Timothy E. Hewett
Keyword(s):  
Discriminant Analysis ◽  
Linear Discriminant Analysis ◽  
Knee Injury ◽  
Factor Model ◽  
Tibial Rotation ◽  
Linear Discriminant ◽  
Mechanical Impact ◽  
Knee Kinetics ◽  
Input Variables ◽  
Injury Outcomes

Background: The most commonly damaged structures of the knee are the anterior cruciate ligament (ACL), medial collateral ligament (MCL), and menisci. Given that these injuries present as either isolated or concomitant, it follows that these events are driven by specific mechanics versus coincidence. This study was designed to investigate the multiplanar mechanisms and determine the important biomechanical and demographic factors that contribute to classification of the injury outcome. Hypothesis: Linear discriminant analysis (LDA) would accurately classify each injury type generated by the mechanical impact simulator based on biomechanical input variables (ie, ligament strain and knee kinetics). Study Design: Controlled laboratory study. Methods: In vivo kinetics and kinematics of 42 healthy, athletic participants were measured to determine stratification of injury risk (ie, low, medium, and high) in 3 degrees of knee forces/moments (knee abduction moment, anterior tibial shear, and internal tibial rotation). These stratified kinetic values were input into a cadaveric impact simulator to assess ligamentous strain and knee kinetics during a simulated landing task. Uniaxial and multiaxial load cells and implanted strain sensors were used to collect mechanical data for analysis. LDA was used to determine the ability to classify injury outcome by demographic and biomechanical input variables. Results: From LDA, a 5-factor model (Entropy R2 = 0.26) demonstrated an area under the receiver operating characteristic curve (AUC) for all 5 injury outcomes (ACL, MCL, ACL+MCL, ACL+MCL+meniscus, ACL+meniscus) of 0.74 or higher, with “good” prediction for 4 of 5 injury classifications. A 10-factor model (Entropy R2 = 0.66) improved the AUC to 0.86 or higher, with “excellent” prediction for 5 injury classifications. The 15-factor model (Entropy R2 = 0.85), produced 94.1% accuracy with the AUC 0.98 or higher for all 5 injury classifications. Conclusion: Use of LDA accurately predicted the outcome of knee injury from kinetic data from cadaveric simulations with the use of a mechanical impact simulator at 25° of knee flexion. Thus, with clinically relevant kinetics, it is possible to determine clinical risk of injury and also the likely presentation of singular or concomitant knee injury. Clinical Relevance: LDA demonstrates that injury outcomes are largely characterized by specific mechanics that can distinguish ACL, MCL, and medial meniscal injury. Furthermore, as the mechanics of injury are better understood, improved interventional prehabilitation can be designed to reduce these injuries.

scholarly journals Development of cardiotoxicity model using ligand-centric and receptor-centric descriptors

2020 ◽  
Vol 4 ◽  
pp. 239784732097125
Author(s):  
Chirag N Patel ◽  
Sivakumar Prasanth Kumar ◽  
Rakesh M Rawal ◽  
Manishkumar B Thaker ◽  
Himanshu A Pandya
Keyword(s):  
Discriminant Analysis ◽  
Linear Discriminant Analysis ◽  
Statistical Significance ◽  
Characteristic Curve ◽  
Hybrid Approach ◽  
Classification Model ◽  
Test Set ◽  
Linear Discriminant ◽  
Specificity And Sensitivity ◽  
Physico Chemical

Background: Bioinformatics and statistical analysis have been employed to develop a classification model to distinguish toxic and non-toxic molecules. Aims: The primary objective of this study is to enumerate the cut-off values of various physico-chemical (ligand-centric) and target interaction (receptor-centric) descriptors which forms the basis for classifying cardiotoxic and non-toxic molecules. We also sought correlation of molecular docking, absorption, distribution, metabolism, excretion, and toxicology (ADMET) parameters, Lipinski rules, physico-chemical parameters, etc. of human cardiotoxicity drugs. Methods: A training and test set of 91 compounds were applied to linear discriminant analysis (LDA) using 2D and 3D descriptors as discriminating variables representing various molecular modeling parameters to identify which function of descriptor type is responsible for cardiotoxicity. Internal validation was performed using the leave-one-out cross-validation methodology ensuing in good results, assuring the stability of the discriminant function (DF). Results: The values of the statistical parameters Fisher Discriminant Analysis (FDA) and Wilk’s λ for the DF showed reliable statistical significance, as long as the success rate in the prediction for both the training and the test set attained more than 93% accuracy, 87.50% sensitivity and 94.74% specificity. Conclusion: The predictive model was built using a hybrid approach using organ-specific targets for docking and ADMET properties for the FDA (Food and Drug Administration) approved and withdrawn drugs. Classifiers were developed by linear discriminant analysis and the cut-off was enumerated by receiver operating characteristic curve (ROC) analysis to achieve reliable specificity and sensitivity.

scholarly journals Abnormal visual gain control and excitotoxicity in early-onset Parkinson’s disease Drosophila models

2018 ◽  
Vol 119 (3) ◽  
pp. 957-970 ◽  
Author(s):  
Marc M. Himmelberg ◽  
Ryan J. H. West ◽  
Christopher J. H. Elliott ◽  
Alex R. Wade
Keyword(s):  
Parkinson’S Disease ◽  
Steady State ◽  
Discriminant Analysis ◽  
Linear Discriminant Analysis ◽  
Early Onset ◽  
Gain Control ◽  
Temporal Contrast ◽  
Linear Discriminant ◽  
Abnormal Visual ◽  
Drosophila Models

The excitotoxic theory of Parkinson’s disease (PD) hypothesizes that a pathophysiological degeneration of dopaminergic neurons stems from neural hyperactivity at early stages of disease, leading to mitochondrial stress and cell death. Recent research has harnessed the visual system of Drosophila PD models to probe this hypothesis. Here, we investigate whether abnormal visual sensitivity and excitotoxicity occur in early-onset PD (EOPD) Drosophila models DJ-1αΔ 72, DJ-1βΔ 93, and PINK15. We used an electroretinogram to record steady-state visually evoked potentials driven by temporal contrast stimuli. At 1 day of age, all EOPD mutants had a twofold increase in response amplitudes compared with w̄ controls. Furthermore, we found that excitotoxicity occurs in older EOPD models after increased neural activity is triggered by visual stimulation. In an additional analysis, we used a linear discriminant analysis to test whether there were subtle variations in neural gain control that could be used to classify Drosophila into their correct age and genotype. The discriminant analysis was highly accurate, classifying Drosophila into their correct genotypic class at all age groups at 50–70% accuracy (20% chance baseline). Differences in cellular processes link to subtle alterations in neural network operation in young flies, all of which lead to the same pathogenic outcome. Our data are the first to quantify abnormal gain control and excitotoxicity in EOPD Drosophila mutants. We conclude that EOPD mutations may be linked to more sensitive neuronal signaling in prodromal animals that may cause the expression of PD symptomologies later in life. NEW & NOTEWORTHY Steady-state visually evoked potential response amplitudes to multivariate temporal contrast stimuli were recorded in early-onset PD Drosophila models. Our data indicate that abnormal gain control and a subsequent visual loss occur in these PD mutants, supporting a broader excitotoxicity hypothesis in genetic PD. Furthermore, linear discriminant analysis could accurately classify Drosophila into their correct genotype at different ages throughout their lifespan. Our results suggest increased neural signaling in prodromal PD patients.

scholarly journals Depth-Based Classification for Distributions with Nonconvex Support

2013 ◽  
Vol 2013 ◽  
pp. 1-7 ◽  
Author(s):  
Daniel Hlubinka ◽  
Ondrej Vencalek
Keyword(s):  
Statistical Analysis ◽  
Discriminant Analysis ◽  
Linear Discriminant Analysis ◽  
Classification Problem ◽  
Data Depth ◽  
Misclassification Rate ◽  
Skewed Data ◽  
Weighted Version ◽  
Linear Discriminant ◽  
Normal Populations

Halfspace depth became a popular nonparametric tool for statistical analysis of multivariate data during the last two decades. One of applications of data depth considered recently in literature is the classification problem. The data depth approach is used instead of the linear discriminant analysis mostly to avoid the parametric assumptions and to get better classifier for data whose distribution is not elliptically symmetric, for example, skewed data. In our paper, we suggest to use weighted version of halfspace depth rather than the halfspace depth itself in order to obtain lower misclassification rate in the case of “nonconvex” distributions. Simulations show that the results of depth-based classifiers are comparable with linear discriminant analysis for two normal populations, while for nonelliptic distributions the classifier based on weighted halfspace depth outperforms both linear discriminant analysis and classifier based on the usual (nonweighted) halfspace depth.

Sign in / Sign up

Export Citation Format

Share Document