Principal Component Analysis Neural Network Research Articles

A Novel Method for Multivariant Pneumonia Classification Based on Hybrid CNN-PCA Based Feature Extraction Using Extreme Learning Machine With CXR Images

In this era of COVID19, proper diagnosis and treatment of pneumonia are very important. Chest X-Ray (CXR) image analysis plays a vital role in the reliable diagnosis of pneumonia. An experienced radiologist is required for this. However, even for an experienced radiographer, it is quite challenging and time-consuming to diagnose accurately due to the fuzziness of CXR images. Also, identification can be erroneous due to the involvement of human judgement. Hence, an authentic and automated system can play an important role here. In this era of cutting-edge technology, deep learning (DL) is highly used in every sector. There are several existing methods to diagnose pneumonia but they have accuracy problems. In this study, an automatic pneumonia detection system has been proposed by applying the extreme learning machine (ELM) on the Kaggle CXR images (Pneumonia). Three models have been studied: classification using extreme learning machine (ELM), ELM with a hybrid convolutional neural network-principal component analysis (CNN-PCA) based feature extraction, and CNN-PCA-ELM with the CXR images which are contrast-enhanced by contrast limited adaptive histogram equalization (CLAHE). Among these three proposed methods, the final model provides an optimistic result. It achieves the recall score of 98% and accuracy score of 98.32% for multiclass pneumonia classification. On the other hand, a binary classification achieves 100% recall and 99.83% accuracy. The proposed method also outperforms the existing methods. The outcome has been compared using several benchmarks that include accuracy, precision, recall, etc.

Development of Insert Condition Classification System for CNC Lathes Using Power Spectral Density Distribution of Accelerometer Vibration Signals.

Insert conditions significantly influence the product quality and manufacturing efficiency of lathe machining. This study used the power spectral density distribution of the vibration signals of a lathe machining accelerometer to design an insert condition classification system applicable to different machining conditions. For four common lathe machining insert conditions (i.e., built-up edge, flank wear, normal, and fracture), herein, the insert condition classification system was established with two stages—insert condition modeling and machining model fusion. In the insert condition modeling stage, the magnitude features of the segmented frequencies were captured according to the power spectral density distributions of the accelerometer vibration signals. Principal component analysis and backpropagation neural networks were used to develop insert condition models for different machining conditions. In the machining model fusion stage, a backpropagation neural network was employed to establish the weight function between the machining conditions and insert condition models. Subsequently, the insert conditions were classified based on the calculated weight values of all the insert condition models. Cutting tests were performed on a computer numerical control (CNC) lathe and utilized to validate the feasibility of the designed insert condition classification system. The results of the cutting tests showed that the designed system could perform insert condition classification under different machining conditions, with a classification rate exceeding 80%. Using a triaxial accelerometer, the designed insert condition classification system could perform identification and classification online for four common insert conditions under different machining conditions, ensuring that CNC lathes could further improve manufacturing quality and efficiency in practice.

Prediction and Analysis of Strawberry Moisture Content based on BP Neural Network Model

Background: Moisture content is one of the most important indicators for the quality of fresh strawberries. Currently, several methods are usually employed to detect the moisture content in strawberry. However, these methods are relatively simple and can only be used to detect the moisture content of single samples but not batches of samples. Besides, the integrity of the samples may be destroyed. Therefore, it is important to develop a simple and efficient prediction method for strawberry moisture to facilitate the market circulation of strawberry. Objective: This study aims to establish a novel BP neural network prediction model to predict and analyze strawberry moisture. Methods: Toyonoka and Jingyao strawberries were taken as the research objects. The hyperspectral technology, spectral difference analysis, correlation coefficient method, principal component analysis and artificial neural network technology were combined to predict the moisture content of strawberry. Results: The characteristic wavelengths were highly correlated with the strawberry moisture content. The stability and prediction effect of the BP neural network prediction model based on characteristic wavelengths are superior to those of the prediction model based on principal components, and the correlation coefficients of the calibration set for Toyonaka and Jingyao respectively reached up to 0.9532 and 0.9846 with low levels of standard deviations (0.3204 and 0.3010, respectively). Conclusion: The BP neural network prediction model of strawberry moisture has certain practicability and can provide some reference for the on-line and non-destructive detection of fruits and vegetables.

Application of Convolutional Neural Network-Based Feature Extraction and Data Fusion for Geographical Origin Identification of Radix Astragali by Visible/Short-Wave Near-Infrared and Near Infrared Hyperspectral Imaging.

Radix Astragali is a prized traditional Chinese functional food that is used for both medicine and food purposes, with various benefits such as immunomodulation, anti-tumor, and anti-oxidation. The geographical origin of Radix Astragali has a significant impact on its quality attributes. Determining the geographical origins of Radix Astragali is essential for quality evaluation. Hyperspectral imaging covering the visible/short-wave near-infrared range (Vis-NIR, 380–1030 nm) and near-infrared range (NIR, 874–1734 nm) were applied to identify Radix Astragali from five different geographical origins. Principal component analysis (PCA) was utilized to form score images to achieve preliminary qualitative identification. PCA and convolutional neural network (CNN) were used for feature extraction. Measurement-level fusion and feature-level fusion were performed on the original spectra at different spectral ranges and the corresponding features. Support vector machine (SVM), logistic regression (LR), and CNN models based on full wavelengths, extracted features, and fusion datasets were established with excellent results; all the models obtained an accuracy of over 98% for different datasets. The results illustrate that hyperspectral imaging combined with CNN and fusion strategy could be an effective method for origin identification of Radix Astragali.

A new method to production prediction for the shale gas reservoir

ABSTRACT It is a common phenomenon that shale gas production differs significantly among wells, and it is difficult to find a relationship between engineering variables and gas rate by utilizing a simple plotting of production data against well and treatment variables. Data such as completion method, flow back, and gas production have been collected for 180 shale gas wells. Firstly, the collected data were standardized, and then related or dependent variables were removed by gray correlation analysis. Finally, principal component analysis and BP Neural-network methods are applied separately to analyze the collected data. It is the first-of-its-kind study that was conducted to define the relationship between gas production and engineering parameters for Fuling shale gas reservoir in China. By use of principal component analysis, it is concluded that the number of stages, the lateral length, the distance between clusters, and the average volume of slick-water are the main factors that impact shale gas production. The results showed that stage numbers account for 42.9% of the total variance, while the horizontal length accounts for about 21.4%; therefore, these two factors account for 64.3% among all of the factors. However, the next important factors are the average distance between clusters and the average volume of slick-water, both account for 24.1% of the total variance. The importance of each variable was also defined by the BP Neural-network method. The number of stages and the horizontal length are still the two major factors, and both account for 52%. The volume of slick-water and the proppant size of 40/70 mesh for both ceramic and sand are considered the second important factor and both account for 37%. This implies that shale gas well configuration and treatment parameters are the dominant factors for gas production in shale reservoirs. Therefore, in order to enhance gas production, it is crucial to increase the number of stages and horizontal length firstly and to raise the volume of slick-water by using 40/70 mesh for both ceramic and sand. However, another driver to accelerate production is to decrease the distance of clusters. These measures will be helpful to improve production in the future shale gas development of Fuling reservoir. This study is believed to be the first-of-its-kind to define the relationship between gas production and engineering parameters for Fuling shale gas reservoir in China.

Accurate prediction of glucose concentration and identification of major contributing features from hardly distinguishable near-infrared spectroscopy

In this study, we tackle the accurate prediction of glucose aqueous concentration from hardly distinguishable near-infrared (NIR) spectroscopy. We adopted several machine learning approaches for the spectral analyses and identified important features learned by each model. The models we investigated include Partial Least Squares Regression (PLSR), Support Vector Machine Regression (SVMR), Random Forest Regression (RF), Extra Trees Regression (ETR), eXtreme Gradient Boosting (Xgboost), and hybrid Principal Component Analysis-Neural Network (PCA-NN) methods. From 47 different glucose aqueous concentrations which cover the range of 40–500 mg/dl, we measured 564 near-infrared (NIR) absorbance spectra samples with wavelength range 900 nm–2200 nm. Then the spectra samples were randomly split into 80% for the training set and 20% for the testing set. In our test, we found that the models SVMR, ETR, and PCA-NN reach extremely good performance, which had correlation coefficient R > 0.99 and determination of coefficient R2 > 0.985. To explore the robustness of each machine learning approach, we extracted their high-weighting features and examine their distribution. We found that having large overlapping of the high-weighting features learned by the model when trained by different data sets may be an indication of model stability. In addition, our analysis came up with the essential region of features to disentangle the hardly distinguishable signal. Our study demonstrates a robust machine learning models for the prediction of glucose aqueous concentration in an in-vitro setup using near-infrared spectroscopy.

Modulation format/bit rate recognition based on principal component analysis (PCA) and artificial neural networks (ANNs): erratum

The aim of this erratum is to add an omitted reference.

Nitrate nitrogen concentration detection method based on principal component analysis and BP neural network

Nitrate nitrogen concentration detection method based on principal component analysis and BP neural network

Non-invasive Estimation of Haemoglobin Level Using PCA and Artificial Neural Networks

Objective: Haemoglobin(Hb) measurement is generally performed by the traditional “fingerstick” test i.e., by invasively drawing blood from the body. Although the conventional laboratory measurement is accurate, it has its own limitations such as time delay, inconvenience of the patient, exposure to biohazards and the lack of real-time monitoring in critical situations. Non-invasive Haemoglobin Measurement (SpHb) has gained enormous attention among researches and can provide an earlier diagnosis to polycythemia, anaemia, various cardiovascular diseases, etc. Currently, Photoplethysmograph signal (PPG) is used for measuring oxygen saturation, to monitor the depth of anesthesia, heart rate and respiration monitoring. But through detailed statistical analysis, PPG signal can provide further information about various blood components. Investigation / Methodology: In this paper, an approach for non-invasive measurement of Hb using PPG, Principal Component Analysis (PCA) and Neural Network is proposed. A transmissive type PPG sensor is developed which is interfaced with Crowduino for the acquisition of PPG. From the obtained PPG signal, Principal Components (PC) are extracted. SpHb is predicted followed by the extraction of features from the PC. The analysis involves the SpHb prediction using a single PC, double PC and finally all the three PC. The predicted SpHb is evaluated with Hblab in terms of R-value, Mean Absolute Error, Mean Squared Error and Root Mean Squared Error. Conclusion: An approach for non-invasive measurement of Hb using Principal Components obtained from the PPG signal is discussed. The SpHb value is compared with the Hblab values. Correlation R-value between SpHb and Hblab is 0.77 when three principal components are used. Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) between SpHb and Hblab are 0.3, 0.44 and 0.6633 respectively when SpHb is measured with three principal components. It is evident from the result analysis that SpHb shows the promising result when all the three principal components are used. However, one of the limitations of the work is that the population setting chosen for the work does not include paediatric patients, accurately ill patient, pregnant population and surgical patients. With detailed analysis on a wide range of population setting, Hb prediction using PPG is a promising approach for non-invasive measurement.

A neural-network-based framework for cigarette laser code identification

The identification of cigarette laser codes is important in distinguishing the authenticity of tobacco. However, the existing character recognition methods have limited use in the identification due to the complex background in cigarette images. To address this issue, we propose a novel neural-network-based framework in this paper. Specifically, the framework includes three major steps. Firstly, a principal component analysis neural network is designed for the inclination correction progress to overcome the strong noise interferences. Then a novel algorithm is proposed to adaptively utilize the prior partition information for better character segmentation. Finally, a CNN model is designed to extract irregular features for character identification. By doing this, the proposed framework alleviates the influence of diverse backgrounds and keeps useful features at the same time. Additionally, we give an insight analysis on the character recognition based on the proposed method. The performance of the framework is evaluated on an image set composed of 100 cigarette laser code photos, whose results demonstrate that our framework can bring about 30% improvement in recognition accuracy compared to baseline methods. The good performance indicates a huge potential of our framework on practical applications.

Radar Data Cube Processing for Human Activity Recognition Using Multisubspace Learning

In recent years, radar has been employed as a fall detector because of its effective sensing capabilities and penetration through walls. In this paper, we introduce a multilinear subspace human activity recognition scheme that exploits the three radar signal variables: slow-time, fast-time, and Doppler frequency. The proposed approach attempts to find the optimum subspaces that minimize the reconstruction error for different modes of the radar data cube. A comprehensive analysis of the optimization considerations is performed, such as initialization, number of projections, and convergence of the algorithms. Finally, a boosting scheme is proposed combining the unsupervised multilinear principal component analysis (PCA) with the supervised methods of linear discriminant analysis and shallow neural networks. Experimental results based on real radar data obtained from multiple subjects, different locations, and aspect angles (0°, 30°, 45°, 60°, and 90°) demonstrate that the proposed algorithm yields the highest overall classification accuracy among spectrogram-based methods including predefined physical features, one- and two-dimensional PCA and convolutional neural networks.

A Robust PCA-SURE Thresholding Deep Neural Network Approach for Mental Task Brain Computer Interface

Electroencephalographic (EEG) characteristics, i.e., non-linear structure and nonstationarity, make mental state recognition not a trivial task to various classification models. In this paper, a combined principal component analysis (PCA) – Deep neural network approach is proposed as a robust and effective solution to classifying EEG signals recorded with low-cost and portable recording systems into different mental states towards implementation of a Brain computer interface (BCI) capable of controlling electronic devices. Stein’s unbiased risk estimate (SURE) thresholding – PCA is utilized to obtain spectral features that are most essential for deep neural network classifier to perform at best. The contributions of this paper are three–fold. First, we propose a novel, robust and efficient method that integrates SURE thresholding, PCA and Deep Neural Network (DNN), along with other signal processing and machine learning techniques for mental state recognition. Second, a complete mental-task-based BCI using an appropriate experimental design without any assistant equipment is presented. Third, SURE risk thresholding is utilized and proven to be an effective method to automatically determine the appropriate number of principal components of EEG features returned by performing PCA. Experimental results show that our method outperforms others in an EEG dataset of four subjects with highest classification accuracies on dual and triple mental state task experiments of 96.83% and 76.90%, respectively.

Topological neural network of combined AE and EN signals for assessment of SCC damage

ABSTRACTStress corrosion cracking (SCC) is a common corrosion form that involves undetected premature failures during service life of structures. Since SCC is a combination of electrochemical and mechanical phenomena, coupled acoustic emission and electrochemical noise techniques were proposed to investigate the evolution of SCC damage tomartensitic stainless steel samples. Tests were carried out using a precipitation hardening martensitic stainless steel in an aqueous MgCl2 environment at 100°C, with an applied mechanical stress equal to the 90% of 0.2% yield strength. The synergistic use of the two non-destructive analysis technique was performed using a synchronisation process. The combination of two advanced multivariate analysis techniques (Principal Component Analysis and Self Organising Map neural network) highlighted damage-sensitive features. Characteristic clusters of variables related to specific damage mechanisms were identified discriminating electrochemical activation processes (i.e. pitting), SCC initiation and propagation until final failure phenomena during SCC.

Exploration of stock index change prediction model based on the combination of principal component analysis and artificial neural network

In order to establish an accurate effective stock forecasting model, the principal component analysis (PCA) was first used to analyze the main financial index data of some listed companies and the comprehensive score of evaluation index was obtained in this study. Then, the financial indicator data and the transaction indicator data were simultaneously used as the input variables of the stock price prediction research, three back propagation (BP) neural network algorithms were used for experiment, and its prediction situation was compared. Results show that the BP neural network based on Bayesian regularization algorithm has the highest prediction accuracy and can avoid over-fitting phenomenon in the training process of the model; the error between the predicted value and the actual value is small. Finally, this study constructed a stock price prediction study based on PCA and BP neural network algorithm as well as an investment stock selection strategy based on traditional stock selection analysis method. As a result, the proposed model is proved to be effective.

Prediction of Negative Conversion Days of Childhood Nephrotic Syndrome Based on PCA and BP-AdaBoost Neural Network

The prognosis of childhood nephrotic syndrome directly hinges on the accurate prediction of negative conversion days (NCDs). Therefore, this paper designs a hybrid approach of principal component analysis (PCA) and backpropagation (BP)-adaptive boosting (AdaBoost) neural network (NN), and applies the method to predict the NCDs for children with nephrotic syndrome. Specifically, PCA method was employed for dimension reduction. Six principal components were extracted from multiple physiological features, and taken as input variables of three-tiered neural networks. The boosted predictor of BP-AdaBoost model, together with three predictors of BP NN, support vector machine (SVM) and radial basis function (RBF) NN, was trained for NCDs prediction. The experimental results show that the predictor of BP-AdaBoost NN achieves the mean absolute error of 0.2334, the mean relative error of 3.2789%, the SD of 4.6804 and the RSD of 50.8053% in NCDs prediction, and it outperforms other predictors of BP NN, SVM and RBF NN on both accuracy and precision. Furthermore, comparison experiments are conducted on PCA processed testing data and raw testing data for BP-AdaBoost NN and demonstrate the excellent effect of PCA. The hybrid approach of PCA and BP-AdaBoost NN is simple and reliable for NCDs prediction of childhood nephrotic syndrome, and it can help pediatricians prognose childhood nephrotic syndrome accurately and further provide patients with better care and treatment.

Development and Deployment of Organizational Culture Maturity Model (TMKC) for the Business Excellence Success

The paper describes the development of the maturity-based organisational culture model (TMKC - team bonding-motivation-knowledge management-commitment for improvement). Each level of the model has practices for the improvements. The model deployment in 11 different organisations indicates the earnings before interest, depreciation, taxes and amortisation (EBITDA) are significantly higher. Delphi method-based survey, dipstick survey and correlation analysis were applied to discover the extent of the problem. The cluster analysis, principal components analysis and artificial neural network were used to prioritise organisation-based cultural factors from 13 to 4. The impact of TMKC model was validated through pilots and statistical tests. The higher maturity levels of 4 and 5 have significantly increased business excellence (BE) initiative goal achievement rate from 34% to 86%. The successful TMKC model deployment positively impacts % EBITDA change. The model has been piloted in 11 cases and its finding can change with deployments in varied instances.

Assessing the Forecasting of Comprehensive Loss Incurred by Typhoons: A Combined PCA and BP Neural Network Model

This paper develops a joint model utilizing the principal component analysis (PCA) and the back propagation (BP) neural network model optimized by the Levenberg Marquardt (LM) algorithm, and as an application of the joint model to investigate the damages caused by typhoons for a coastal province, Fujian Province, China in 2005-2015 (latest). First, the PCA is applied to analyze comprehensively the relationship between hazard factors, hazard bearing factors and disaster factors. Then five integrated indices, overall disaster level, typhoon intensity, damaged condition of houses, medical rescue and self-rescue capability, are extracted through the PCA; Finally, the BP neural network model, which takes the principal component scores as input and is optimized by the LM algorithm, is implemented to forecast the comprehensive loss of typhoons. It is estimated that an average annual loss of 138.514 billion RMB occurred for 2005-2015, with a maximum loss of 215.582 in 2006 and a decreasing trend since 2010 though the typhoon intensity increases. The model was validated using three typhoon events and it is found that the error is less than 1%. These results provide information for the government to increase medical institutions and medical workers and for the communities to promote residents’ self-rescue capability.

Recognition Method of Drinking-driving Behaviors Based on PCA and RBF Neural Network

Drinking-driving behaviors are important causes of road traffic injuries, which are serious threats to the lives and property of traffic participants. Therefore, reducing the occurrences of drinking-driving behaviors has become an important problem of traffic safety research. Forty-eight male drivers and six female drivers who could drink moderate alcohol were chosen as participants. The drivers’ physiological data, operation behavior data, car running data, and driving environment data were collected by designing various virtual traffic scenes and organizing drivers to conduct driving simulation experiments. The original variables were analyzed by the Principal Component Analysis (PCA), and seven principal components were extracted as the input vector of the Radial Basis Function (RBF) neural network. The principal component data was used to train and verify the RBF neural network. The Levenberg-Marquardt (LM) algorithm was chosen to train the parameters of the neural network and build a drinking-driving recognition model based on PCA and RBF neural network to realize an accurate recognition of drinking-driving behaviors. The test results showed that the drinking-driving recognition model based on PCA and RBF neural network could identify drinking drivers accurately during driving process with a recognition accuracy of 92.01%, and the operation efficiency of the model was high. The research can provide useful reference for prevention and treatment of drinking and driving and traffic safety maintenance.

Prediction of wind drift and evaporation losses of a sprinkler irrigation system using principal component analysis and artificial neural network technique

Principal component analysis was merged with the artificial neural network (ANN) technique to predict wind drift and evaporation losses (WDEL) from a sprinkler irrigation system. For this purpose, field experiments were conducted to determine WDEL under different conditions. Data from field experiments and previous studies were used as sample data to train the ANN model. Three models were developed to predict WDEL. In the first model (ANN1), 9 neurons (riser height, main nozzle diameter, auxiliary nozzle diameter, discharge rate of the main nozzle, discharge rate of the auxiliary nozzle, operating pressure, wind speed, air temperature and relative humidity) were used as the input layer. In the second model (ANN2), 7 neurons (riser height, operating pressure, wind speed, air temperature and relative humidity, diameter ratio and discharge ratio) were used as the input layer. The third model (ANN3) used a multivariate technique (PC1, PC2, and PC3). Results revealed that the ANN3 model had the highest coefficient of determination (R2 = 0.8349). The R2 values for the ANN1 and ANN2 models were 0.7792 and 0.4807, respectively. It can be concluded that the ANN3 model has the highest predictive capacity.

A Hybrid Approach for Breast Cancer Classification and Diagnosis

Feature selection in breast cancer disease important and risky task for further analysis. Breast cancer is the second leading reason for death among the women. Cancer starts from breast and spread to other part of the body. People are unable to identify their disease before it become dangerous. It can be cured if the disease identified at early stage. Accurate classification of benign tumours can avoid patients undergoing unnecessary treatments. Data Analytics and machine learning methods provides framework for prognostic studies by errorless classification of data instances into relevant based on the cancer severity. In this study we have purposed a prediction model by combining artificial intelligent based learning technique with multivariate statistical method. For automation of the diagnosis process data mining plays an significant role. The data sets available in different repositories are noisy in nature. This study suggests a hybrid feature selection method to be used with PCA (Principal Component Analysis) and Artificial Neural Network (ANN). Preprocessing of data and extracting the most relevant features done by PCA. The proposed algorithm is tested by applying it on Wisconsin Breast Cancer Dataset from UCI Repository of Machine Learning Databases. In classification phase 10 fold cross validation was used. The suggested algorithm was measured against different classifier algorithms on the same database. The evaluation results of the algorithm proposed have achieved better accuracy with sensitivity and F measure comparison with others and by enhancing this concept we can provide a future scope to produce sophisticated learning models for diagnosis.