Hybrid Principal Component Research Articles

THE EFFECT OF ANTHROPOMETRIC PARAMETERS ON THE SEAT TRANSMISSIBILITY DURING WHOLE-BODY VIBRATION: A HYBRID MODELING APPROACH BASED ON ARTIFICIAL NEURAL NETWORKS AND PRINCIPAL COMPONENT ANALYSIS

The modeling of the seat transmissibility is necessary to advance the understanding of the dynamic interactions between compliant seats and occupants. Within this investigation, an optimized artificial neural network (ANN) model was employed to clarify contributions associated with anthropometric parameters to the seat transmissibility. Anthropometric parameters underwent dimensionality reduction through the principal component analysis, and resultant principal components served as input features for the ANN model. Additionally, the ANN structure’s weights and biases values were adjusted using the genetic algorithm (GA), resulting in a PCA-GA-ANN model for the prediction of seat transmissibilities. The results indicated root mean square error (RMSE) values for predicting vertical in-line and horizontal cross-axis transmissibilities from the developed model were 0.061 and 0.055, respectively, demonstrating superior effectiveness in the prediction error and trends when compared with both the ANN and GA-ANN models. The seat transmissibility predicted from the PCA-GA-ANN model exhibited resonance behaviors similar to that observed in the whole-body vibration test. The sensitivity analysis showed that the subject’s age was the most predominant anthropometric parameter for the prediction, followed by gender and body mass index. The ANN model optimized with principal component analysis (PCA) and GA effectively eliminates the redundant information of anthropometric parameters, enhancing the generalization of the seat transmissibility prediction.

Geographical origin identification of Khao Dawk Mali 105 rice using combination of FT-NIR spectroscopy and machine learning algorithms

The mislabelled Khao Dawk Mali 105 rice coming from other geographical region outside the Thung Kula Rong Hai region is extremely profitable and difficult to detect; to prevent retail fraud (that adversely affects both the food industry and consumers), it is vital to identify geographical origin. Near infrared spectroscopy can be used to detect the specific content of organic moieties in agricultural and food products. The present study implemented the combinatorial method of FT-NIR spectroscopy with chemometrics to identify geographical origin of Khao Dawk Mali 105 rice. Rice samples were collected from 2 different region including the north and northeast of Thailand. NIR spectra data were collected in range of 12,500 – 4,000 cm−1 (800–2,500 nm). Five machine learning algorithms including linear discriminant analysis (LDA), partial least squares discriminant analysis (PLS-DA), C-support vector classification (C-SVC), backpropagation neural networks (BPNN), hybrid principal component analysis-neural network (PC-NN) and K-nearest neighbors (KNN) were employed to classify NIR data of rice samples with full wavelength and selected wavelength by Extremely Randomized Trees (Extra trees) algorithm. Based on the findings, geographical origin of rice could be specified quickly, cheaply, and reliably using combination of NIRS and machine learning. All models creating by full wavelength and selected wavelength exhibited accuracy between 65 and 100 % for identifying geographical region of rice. It was proven that NIR spectroscopy may be used for the quick and non-destructive identification of geographical origin of Khao Dawk Mali 105 rice.

Contextual Nonlinear Typological Analysis as a Tool for the Formation and Implementation of Regional Economic Policy

The article outlines the main theoretical and methodological positions on the formation and implementation of regional economic policy in relation to technological innovations. It is based on contextual typological analysis using economic and mathematical modeling, where the key role is played by the hybrid principal component method and scaling technology. Official statistics data are used as information and analytical support. The concept of network rating for spatial economic systems is introduced and the analysis of regional systems is carried out. A convenient visualization system for analyzing and making management decisions based on statistical information is proposed.

Software Effort Estimation Using Stacked Ensemble Technique and Hybrid Principal Component Regression and Multivariate Adaptive Regression Splines

Software Effort Estimation Using Stacked Ensemble Technique and Hybrid Principal Component Regression and Multivariate Adaptive Regression Splines

An augmented hybrid multibaseline and referenceless MR thermometry motion compensation algorithm for MRgHIFU hyperthermia.

A hybrid principal component analysis and projection onto dipole fields (PCA-PDF) MR thermometry motion compensation algorithm was optimized with atlas image augmentation and validated. Experiments were conducted on a 3T Philips MRI and Profound V1 Sonalleve high intensity focused ultrasound (high intensity focused ultrasound system. An MR-compatible robot was configured to induce motion on custom gelatin phantoms. Trials with periodic and sporadic motion were introduced on phantoms while hyperthermia was administered. The PCA-PDF algorithm was augmented with a predictive atlas to better compensate for larger sporadic motion. During periodic motion, the temperature SD in the thermometry was improved from to C with both the original and augmented PCA-PDF application. For large sporadic motion, the augmented atlas improved the motion compensation from the original PCA-PDF correction from to C. The PCA-PDF algorithm improved temperature accuracy to <1°C during periodic motion, but was not able to adequately address sporadic motion. By augmenting the PCA-PDF algorithm, temperature SD during large sporadic motion was also reduced to <1°C, greatly improving the original PCA-PDF algorithm.

BivalveNet: A hybrid deep neural network for common cockle (Cerastoderma edule) geographical traceability based on shell image analysis

Bivalve traceability is a major concern. It is of utmost importance to develop tools that allow providing important information to the consumer, not only on the origin of the product but also on its sustainability and safety, due to the harvest restrictions imposed by regulatory entities. This study evaluated the application of computer vision machine learning technologies for efficiently discriminating cockle harvesting origin based on shell geometric and morphometric analysis, improving the traceability methodologies in these organisms, and highlighting the potential of these low-cost techniques as a reliable traceability tool. Thirty Cerastoderma edule samples were collected along the five locations in Atlantic West and South Portuguese coast with individual images processed using lazysnapping segmentation, spectro-textural-morphological phenotype extraction, and feature selection through hybrid Principal Component Analysis and Neighborhood Component Analysis which resulted in R, a*, b*, entropy, and diameter. Three approaches of traceability models were developed and tested: pre-trained networks (EfficientNet-Bo, ResNet101, MobileNetV2, InceptionV3) with numerical inputs (Approach 1), image-based pre-trained networks (Approach 2), and hybrid deep neural networks of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Bidirectional LSTM (BiLSTM) (Approach 3). Based on the test results, Approach 3 with GRU-LSTM-BiLSTM sequence exhibited the highest accuracy (96.91%) and sensitivity (96%) among the other thirteen machine learning models, hence, named as BivalveNet. Comparing the attained accuracy from the BivalveNet to other mollusc traceability studies, it was observed that an efficiency close to the attained using standard destructive, time-consuming, and expensive techniques, making BivalveNet a highly advantageous approach for common cockle geographical traceability studies, available for application to other bivalve species.

Aquaphotomics based screening of tomato powder extracts reveals susceptibility to bulking and coloring agents

Gaining prominence as indispensable culinary ingredients, tomato powders aren't exempt from fraudulent practices. While the complexity of conventional authentication methods hinders their adoption as routine assessment tools, our study examines the efficiency of aquaphotomics as a potential alternative, that entails limited preparation steps and involves no solvent extractions. Tested across three tomato varieties, aqueous blends comprising concentrations in the range [0%–20%] of two adulterant types (bulking agent, coloring agent), added separately or concurrently, were considered. Upon scanning with an XDS Near infrared spectrometer operating in transmission mode, the data was subjected to chemometrics based analysis over the 1st and 2nd water overtone ranges. While attempting to perform classification based on the level of adulteration, running hybrid principal component analysis and linear discriminant analysis (PCA-LDA) qualitative modeling on single adulterant mixtures allowed for recognition and prediction rates varying from [97.76%–100%] and [88.59%–100%], respectively. Slightly lower accuracies were obtained when coupling the bulking and coloring agent. Partial least squares regression (PLSR) models gave good estimates of the extent of adulteration, with respective R2 and RMSECV values reaching at best, 0.98 and 0.45%w/w. The aquagrams revealed distinctive spectral patterns highlighting the gradual level of adulteration. Assuredly, these findings ascertain the contributions aquaphotomics promises in ensuring the authenticity of tomato powders, when analyzed in their liquid state.

A supplier selection & order allocation planning framework by integrating deep learning, principal component analysis, and optimization techniques

In supply chain management, selection of suitable suppliers and allocating corresponding orders are two essential strategic decisions. Making these decisions is a complex process due to some uncertain parameters, such as future demand. This study proposes a three-stage solution framework to solve problems with supplier selection & order allocation planning. In Stage 1, a new modified relational deep learning forecasting technique is developed to forecast the demands of products. In this part, the efficiency of this modified technique is compared with two well-known forecasting techniques, namely Seasonal Auto-Regressive Integrated Moving Average (SARIMA) and Light-Gradient Boosted Machine (LGBM). In Stage 2, a new hybrid principal component analysis method is used to generate suppliers’ weights. The results from Stages 1 and 2 are used in the multiple objectives optimization model which is developed in Stage 3. The hybrid method is used to derive a set of efficient solutions. The developed framework is discussed using a real dataset from the Canadian meat industry. The results of forecasting models show that the developed deep learning network can reduce the forecasting error by 55.42% when compared to the SARIMA method, and 13.1% when compared to the LGBM method. It is also observed that the consideration of inter-product correlation functions can change the selected suppliers and the corresponding orders.

Monitoring of salinity of water on the THA CHIN River basin using portable Vis-NIR spectrometer combined with machine learning algorithms

The goal of this work is to study the alternative practices for monitoring the salinity of water using a combination of portable vis-NIR spectrometer and machine learning approachs. Along 80 km of the Tha Chin River basin from the Gulf of Thailand, the data of salinity and NIR spectrum were collected during winter and summer seasons of Thailand. Salinity of water samples was measured by using a handheld electrical conductivity meter and NIR spectra was recorded with portable FQA-NIR GUN in the wavelength range of 600 to 1100 nm. The 10 machine learning models including partial least square regression (PLS), support vector machine (SVR), decision tree (DT), random forest (RF), adaptive boosting (AB), gradient boosting (GB), bagging meta-estimator (BME), extremely randomized trees (ERT), backpropagation neural networks (BPNN) and hybrid principal component analysis-neural network (PCNN) were applied to train the NIRs models for predicting salinity. All machine learning algorithms showed good prediction results which Rp2 values were higher than 0.84. The models built by tree-based algorithms (DT, RF, AB, GB, BME and ERT) displayed higher performances of calibration set and prediction set than those of PLS, SVM, BPNN and PCNN. Among these, the ERT algorithm showed the best performance Rp2 of 0.97, RMSEP of 0.41 g/L and RPD of 6.00. It was shown that NIR spectroscopy coupled with machine learning could be an alternative simpler way for predicting salinity of water.

Hybrid Principal Component Analysis Denoising Enables Rapid, Label-Free Morpho-Chemical Quantification of Individual Nanoliposomes.

Laser tweezers Raman spectroscopy enables multiplexed, quantitative chemical and morphological analysis of individual bionanoparticles such as drug-loaded nanoliposomes, yet it requires minutes-scale acquisition times per particle, leading to a lack of statistical power in typical small-sized data sets. The long acquisition times present a bottleneck not only in measurement time but also in the analytical throughput, as particle concentration (and thus throughput) must be kept low enough to avoid swarm measurement. The only effective way to improve this situation is to reduce the exposure time, which comes at the expense of increased noise. Here, we present a hybrid principal component analysis (PCA) denoising method, where a small number (∼30 spectra) of high signal-to-noise ratio (SNR) training data construct an effective principal component subspace into which low SNR test data are projected. Simulations and experiments prove the method outperforms traditional denoising methods such as the wavelet transform or traditional PCA. On experimental liposome samples, denoising accelerated data acquisition from 90 to 3 s, with an overall 4.5-fold improvement in particle throughput. The denoised data retained the ability to accurately determine complex morphochemical parameters such as lamellarity of individual nanoliposomes, as confirmed by comparison with cryo-EM imaging. We therefore show that hybrid PCA denoising is an efficient and effective tool for denoising spectral data sets with limited chemical variability and that the RR-NTA technique offers an ideal path for studying the multidimensional heterogeneity of nanoliposomes and other micro/nanoscale bioparticles.

Hybrid Feature Extraction with Ensemble Classifier for Brain Tumor Classification

Brain tumor is one of the most severe nervous system disorders affecting the health of humans and critically it will lead to death. The most elevated disease that causes a major death rate is Glioma, i.e. a primary intracranial tumor. One of the widely used techniques in medical imaging is Magnetic Resonance Imaging (MRI) which turned out as the principle diagnosis model for the analysis of glioma and its treatment. However, the brain tumor segmentation and classification process are more complicated problems to execute. This paper intends to introduce a novel brain tumor classification model that includes four major phases: (i) Pre-processing (ii) Segmentation (iii) Brain Feature extraction (iv) Brain tumor Classification. Initially, the input image is subjected to the pre-processing phase, in which the image is pre-processed under a certain process. The pre-processed images are then subjected to the segmentation phase, which is carried out by the k-means clustering. Subsequently, the segmented images are subjected to the brain feature extraction phase, in which the features are extracted using the hybrid Principal Component Analysis (PCA)-GIST feature extraction method. Then, these features are given as the input to the classification process, where the ensemble classifier is exploited for the same. Moreover, the proposed ensemble technique includes k-Nearest Neighbor (k-NN), Optimized Neural Network (NN), Random Forest (RF), and Support Vector Machine (SVM). To precisely detect the tumor classification, the NN training is performed using Elephant Herding Optimization with mutation operations (EHOMO) Algorithm via selecting the optimal weights.

The Functional Regression With Reconstructed Functions From Hybrid Principal Components Analysis: With EEG-fMRI Application

Objective: In this article, we reconstruct the hybrid data with hybrid principal component analysis (HPCA) as a feature extraction step and model them with functional regression as a modeling step, and comparing models and choose the best number of HPCA based on the prediction accuracy as an evaluation step. Method: We decompose the hybrid data to the eigencomponents with HPCA. The reconstructed data from HPCA were divided into the training and testing dataset. The function-on-function signal compression and scalar-on-function regressions were used. Three simulation scenarios and their applications in the neuroimaging datasets (EEG-fMRI) were studied. The number of HPCA was selected with the mean squared prediction error (MSPE). Result: The simulation shows that the raw data, reconstructed from the first and all HPCAs for the training dataset has median MSPE 0.1001, 0.0028, and 0.0174 respectively, and for the testing, the dataset has 0.3207, 0.1118, and 0.2484 respectively. The EEG-fMRI suggests that in both auditory and visual tasks and standard and target stimuli for different regions of the brain the first HPCA has the smallest MSPE. Conclusions: We conclude this method improves the prediction accuracy of the experiments with the EEG datasets. And we recommend that instead of using the functional PCA on the desired dimension, reconstruct the data with HPCA and average it on the other two dimensions for functional regression models.

Development of variational quantum deep neural networks for image recognition

Parametrized quantum circuits are widely used for supervised learning tasks such as image classification in the noisy intermediate scale quantum era. However, normally, it can only handle low-dimensional data. This study presented a variational quantum deep neural network (VQDNN) model for various scale image recognition tasks. Three classifiers were designed to verify the classification performance of the proposed VQDNN model. In the first classifier, to accommodate the limitations of qubits in both simulation hardware and real quantum hardware – we adopted hybrid principal component analysis – VQDNN architecture. Moreover, the amplitude encoding scheme and the rotation angle coding scheme were employed in the subsequent two classifiers to handle large-size images. Finally, we used the classical neural network and VQDNN model to conduct a comparative experiment of the ten-label classification learning task on the same dataset. The quantum numerical experiment was implemented on two benchmark datasets: the MNIST and UCI databases of handwritten digits. The simulation results showed that the proposed VQDNN classified the two datasets with an accuracy of 100% for the two-class classification task, while the UCI dataset has an accuracy of 90.87% for the ten-label classification task. The proposed VQDNN achieved better classification accuracy than the original classical neural network even under a limited number of qubits available in current hardware, indicating the promising application potential of VQDNN in image recognition.

Multilevel hybrid principal components analysis for region-referenced functional electroencephalography data.

Electroencephalography experiments produce region-referenced functional data representing brain signals in the time or the frequency domain collected across the scalp. The data typically also have a multilevel structure with high-dimensional observations collected across multiple experimental conditions or visits. Common analysis approaches reduce the data complexity by collapsing the functional and regional dimensions, where event-related potential (ERP) features or band power are targeted in a pre-specified scalp region. This practice can fail to portray more comprehensive differences in the entire ERP signal or the power spectral density (PSD) across the scalp. Building on the weak separability of the high-dimensional covariance process, the proposed multilevel hybrid principal components analysis (M-HPCA) utilizes dimension reduction tools from both vector and functional principal components analysis to decompose the total variation into between- and within-subject variance. The resulting model components are estimated in a mixed effects modeling framework via a computationally efficient minorization-maximization algorithm coupled with bootstrap. The diverse array of applications of M-HPCA is showcased with two studies of individuals with autism. While ERP responses to match vs mismatch conditions are compared in an audio odd-ball paradigm in the first study, short-term reliability of the PSD across visits is compared in the second. Finite sample properties of the proposed methodology are studied in extensive simulations.

Covariate-adjusted hybrid principal components analysis for region-referenced functional EEG data.

Electroencephalography (EEG) studies produce region-referenced functional data via EEG signals recorded across scalp electrodes. The high-dimensional data can be used to contrast neurodevelopmental trajectories between diagnostic groups, for example between typically developing (TD) children and children with autism spectrum disorder (ASD). Valid inference requires characterization of the complex EEG dependency structure as well as covariate-dependent heteroscedasticity, such as changes in variation over developmental age. In our motivating study, EEG data is collected on TD and ASD children aged two to twelve years old. The peak alpha frequency, a prominent peak in the alpha spectrum, is a biomarker linked to neurodevelopment that shifts as children age. To retain information, we model patterns of alpha spectral variation, rather than just the peak location, regionally across the scalp and chronologically across development. We propose a covariate-adjusted hybrid principal components analysis (CA-HPCA) for EEG data, which utilizes both vector and functional principal components analysis while simultaneously adjusting for covariate-dependent heteroscedasticity. CA-HPCA assumes the covariance process is weakly separable conditional on observed covariates, allowing for covariate-adjustments to be made on the marginal covariances rather than the full covariance leading to stable and computationally efficient estimation. The proposed methodology provides novel insights into neurodevelopmental differences between TD and ASD children.

Straightening skewed markets with an index tracking optimizationless portfolio

Among professionals and academics alike, it is well known that active portfolio management is unable to provide additional risk-adjusted returns relative to their benchmarks. For this reason, passive wealth management has emerged in recent decades to offer returns close to benchmarks at a lower cost. In this article, we first refine the existing results on the theoretical properties of oblique Brownian motion. Then, assuming that the returns follow skew geometric Brownian motions and that they are correlated, we describe some statistical properties for the \emph{ex-post}, the \emph{ex-ante} tracking errors, and the forecasted tracking portfolio. To this end, we develop an innovative statistical methodology, based on a benchmark-asset principal component factorization, to determine a tracking portfolio that replicates the performance of a benchmark by investing in a subset of the investable universe. This strategy, named hybrid Principal Component Analysis (hPCA), is applied both on normal and skew distributions. In the case of skew-normal returns, we propose a framework for calibrating the model parameters, based on the maximum likelihood estimation method. For testing and validation, we compare four alternative models for index tracking. The first two are based on the hPCA when returns are assumed to be normal or skew-normal. The third model adopts a standard optimization-based approach and the last one is used in the financial sector by some practitioners. For validation and testing, we present a thorough comparison of these strategies on real-world data, both in terms of performance and computational efficiency. A noticeable result is that, not only, the suggested lean PCA-based portfolio selection approach compares well versus cumbersome algorithms for optimization-based portfolios, but, also, it could provide a better service to the asset management industry.

The Comparison between Visually and Auditory Oddball Tasks in the EEG Experiment with Healthy Subjects

Purpose: We compare the two oddball tasks for auditory and visually in the healthy subjects with the EEG experiment. In this regard, we compare three different dimensions of the EEG, including regional, longitudinal, and functional effects. The regional dimension defines as the region of the brain, the longitudinal dimension is the repetition of the stimuli, and the functional dimension is the whole curve of ERP, evoked related potential, instead of the only amplitude and latency of it. We also study the trial effects for each stimulus in each task. Materials and Methods: Seventeen healthy patients, 11 of whom are males, have participated in this study. The dataset was downloaded from the internet. The EEG dataset was cleaned. There are three trials and each has a total of 25 target stimuli per group. The functional data analysis, Hybrid principal component analysis, were used to estimate three effects among two different visually and auditory oddball tasks. The 95% Bayesian credible sets were estimated with three trials of each stimulus and each group. The Generalized Additive Mixed Model was used for studying the effect of trials as random effects. Results: The first eigenfunctions of functional and longitudinal dimensions and first and second eigenvector of regional dimensions were estimated. The 95% Bayesian credible sets indicated that the variability between trials exists and is different for each stimulus and each target. The GAMM model indicates that the interaction effects in the functional dimension are statistically significant for both tasks. The interaction effect in the longitudinal dimension is not statistically significant in the Auditory Task and is statistically significant in the visual task. The brain regions are statistically significant and the main effect of the target stimuli is significant for auditory (P-Value: &lt;0.00) and is not significant for the visual tasks (P-Value: &gt;0.06). The interaction effect for target stimuli and the region of the brain is only significant for Occipital, Parietal, and Right Temporal (P-values &lt; 0.05) for the auditory task. The random effects of trials are statistically significant in most of them. Conclusion: The functional data analysis provides many statistical methods to analyze the EEG dataset. The HPCA can capture the functional-longitudinal and regional dimensions and we also study the new dimension, trials with GAMM. The different regions of the brain have not the same activity in these two tasks. The repeating of the stimuli has a positive effect on complex tasks. The between trial's variability are statistically significant, and we conclude to study this effect to show the stability of the trials.

Prediction of dynamic modulus of asphalt concrete using hybrid machine learning technique

ABSTRACT Due to the issues like operational difficulties, and extensive resource requirements at pavement design stage, dynamic modulus () is estimated using predictive models. However, these models suffer from issues like systematic bias, prediction accuracy, and extensive testing requirements. This study presents a novel hybrid Principal Component Analysis (PCA) – Gene Expression Programming (GEP) approach to predict the of asphalt concrete. The database developed during NCHRP 9-19 study was used for developing this methodology. The information of all properties (i.e. variables) was used as input. PCA helped in removing the redundancy at the input stage while reducing the dimensionality. The extracted principal components (PC's) were used to develop first set of predictive models. The second set of predictive models were developed using the parameters mostly contributing to the individual PC's. Comparison of these two sets indicated that predictive model obtained using variables as direct input resulted in improved accuracy. Comparison of this finalized model with the existing regression-based equations using goodness of fit indicators indicated that proposed hybrid model offers efficient and accurate alternative. The proposed model has flexibility to be used with any new database with recalibration.

Traffic Crash Severity Prediction-A Synergy by Hybrid Principal Component Analysis and Machine Learning Models.

The accurate prediction of road traffic crash (RTC) severity contributes to generating crucial information, which can be used to adopt appropriate measures to reduce the aftermath of crashes. This study aims to develop a hybrid system using principal component analysis (PCA) with multilayer perceptron neural networks (MLP-NN) and support vector machines (SVM) in predicting RTC severity. PCA shows that the first nine components have an eigenvalue greater than one. The cumulative variance percentage explained by these principal components was found to be 67%. The prediction accuracies of the models developed using the original attributes were compared with those of the models developed using principal components. It was found that the testing accuracies of MLP-NN and SVM increased from 64.50% and 62.70% to 82.70% and 80.70%, respectively, after using principal components. The proposed models would be beneficial to trauma centers in predicting crash severity with high accuracy so that they would be able to prepare for appropriate and prompt medical treatment.

A novel PCA–whale optimization-based deep neural network model for classification of tomato plant diseases using GPU

The human population is growing at a very rapid scale. With this progressive growth, it is extremely important to ensure that healthy food is available for the survival of the inhabitants of this planet. Also, the economy of developing countries is highly dependent on agricultural production. The overall economic balance gets affected if there is a variance in the demand and supply of food or agricultural products. Diseases in plants are a great threat to the yield of the crops thereby causing famines and economy slow down. Our present study focuses on applying machine learning model for classifying tomato disease image dataset to proactively take necessary steps to combat such agricultural crisis. In this work, the dataset is collected from publicly available plant–village dataset. The significant features are extracted from the dataset using the hybrid-principal component analysis–Whale optimization algorithm. Further the extracted data are fed into a deep neural network for classification of tomato diseases. The proposed model is then evaluated with the classical machine learning techniques to establish the superiority in terms of accuracy and loss rate metrics.