Intelligent Classification Research Articles

A new intelligent sunflower optimization based explainable artificial intelligence approach for early‐age concrete compressive strength classification and mixture design of RAC

AbstractThis study aims to develop a new artificial intelligence model that can produce explainable rules to predict the mix design and early‐age concrete compressive strength classes of recycled aggregate concrete (RAC). Unlike other black‐box machine learning methods and rule‐based algorithms, the study relies on a metaheuristic mechanism for explainability. This metaheuristic mechanism is not used for a traditional parameter optimization, but to automatically extract interpretable and interpretable rules from the experimental data. In the study, 30 series of RACs are produced, and the samples’ 1‐ and 3‐day early‐age concrete compressive strength values are determined. Concretes produced using these strength values are classified. The labels defined for each concrete class are Class A (C 8/10), Class B (C 12/15), Class C (C 16/20), and Class D (C 20/25). The proposed intelligent classification model that consists of rule set automatically produces interpretable and comprehensible rules from data to determine the early‐age concrete compressive strength class and RAC mix amounts. In addition, the proposed method eliminates the black‐box disadvantages of classical machine learning methods with its explainability and interpretability feature. The sunflower optimization algorithm is adapted as the metaheuristic mechanism and a special fitness function and representative solution form are developed for automatic extraction of high‐quality comprehensible rules by simultaneously optimizing many different metrics. This paper is the first interpretable and comprehensible artificial intelligence model attempt used for early‐age compressive strength classification and mixture design of recycled aggregate concrete by balancing and optimizing both the accuracy and explainability. Proposed explainable intelligent classification model is tested against both well‐known state‐of‐the‐art machine learning algorithms and standard rule‐based methods on the produced real data. Promising results in terms of accuracy, precision, recall are obtained along with the explainability feature.

Design and Experimental Research of Intelligent Inspection and Classification System for Yuba Skin Quality

At present, the surface quality of Yuba skin is determined by sensory methods. In order to realize the intelligent classification detection of Yuba skin quality, this study designed a system that automatically determines the quality of Yuba skin surfaces based on image processing and support vector machine (SVM) approaches. Specifically, the system uses image preprocessing to extract the grayscale eigenvalues, gray level co-occurrence matrix (GLCM) eigenvalues, and gray level run length matrix (GLRLM) eigenvalues of the sample image and uses them as input values for a quality grading system. Through model evaluation of three classification models, the SVM classification model was selected according to the evaluation results, and different kernel functions were used in the model for sample training. Based on Matlab, the quality grading software of Yuba skin was developed and designed. Intelligent detection and grading were realized through the radial basis kernel function support vector machine (RBF-SVM) grading model. The best penalty factor (c = 3.50) and kernel parameter value (g = 0.98) were obtained through cross-validation. The accuracy of the model was 95.31% and 94.16% for the training and test sets, respectively. The grading accuracy of the RBF-SVM grading system was 93.56%, and the error was less than 5% compared with the traditional sensory method of grading; thus, the quality classification method based on the SVM classification system for Yuba skin is feasible and can be used for quality detection.

Convolution neural network fusion high power halogen lamp induced thermal-wave diffusion multidimensional features: An intelligent defect identification approach for aviation honeycomb sandwich composites

HSCs defects may have a serious impact on the mechanical properties of the material. Accurate and reliable non-destructive defect detection of materials can reduce safety hazards. Therefore, a reliable technology for HSCs defect detection is needed. In this present study, a new non-destructive detection method that integrates a convolutional neural network (CNN) and high-power halogen lamp-induced thermal-wave diffusion multidimensional features were introduced to realize the intelligent classification and quantitative characterization of honeycomb sandwich composites (HSCs) defects. Firstly, HSC specimens with delamination defects were detected by the halogen lamp-induced lock-in thermography (LIT) algorithm, and feature images were extracted. After image pre-processing, the images are used as defect feature datasets. Furthermore, combining the four-layer feature pyramid structure and Transformer, a CNN model named YOLOLIT is designed to achieve intelligent classification and identification of defects and quantitative characterization of defect sizes. Subsequently, the optimal hyperparameter combination is obtained by iterating the hyperparameters of the neural network using a genetic algorithm. In a complex environment, accurate recognition of circular defects with a minimum diameter of Φ6mm is achieved, and the mAP of up defect is 90.2%, down defect is 86.2%, remove defect is 72.9%, and the average recognition speed of 0.2 s/image is achieved. Finally, the relation between defect size and detection error as well as that between LIT frequency and detection error is studied. The detection error range and optimal LIT frequency are obtained. The experimental results illustrated this method may be a potential infrared non-destructive defect detection solution.

Research on Key Point Acquisition and Intelligent Classification Technology

By classifying sensitive information, core technologies can be protected, ensuring information not accessed and used by unauthorized personnel, and preventing information from leakage, theft, and abuse. The effective classification work is able to establish an integral information protection mechanism, improve the level of information security, and protect the interests of all parties. In order to manage classified documents better, this paper proposes a multi-level processing method of dynamic key point acquisition and intelligent key point determination, and deep learning models are used to determine the level of document classification comprehensively. The main work of the paper is as follows: using Hidden Markov Model for part of speech tagging in stuttering segmentation algorithm, implementing word vectors by Word2Vec framework, and using CBOW model to predict center words based on contextual word information, and training the model through BERT to determine the classification level of the text automatically.

Dynamic Incipient Fault Forecasting for Power Transformers Using an LSTM Model

Dissolved gas analysis (DGA) is a traditional approach for power transformer fault diagnostics based on measurement of gas contamination. Hydrocarbon gases generated and dissolved in transformer oil during operation can increase in density as fault conditions predominate. Critical determination of gas concentration changes and assessment trending of dissolved gases for fault prediction and prevention of transformer damage is essential. In this paper, a dynamic fault prediction approach is proposed using a long short-term memory (LSTM) model with intelligent classification to determine the running state of a transformer for prediction and avoidance of potential transformer damage. In the paper, the LSTM model processed DGA data collected from real on-site transformer field measurements and predicts future dissolved gas concentrations in time sequence. Four artificial intelligence diagnostic models (support vector machine (SVM), K-nearest neighbors (KNN), decision tree, and artificial neural network (ANN)) were rendered and used for comparative fault prediction assessment. By comparing experimental results from the different LSTM-based models, this paper asserts that the LSTM-KNN model provides the highest and most reliable prediction accuracy for power transformers.

Infant Posture Analyzer and Wet Diaper System

Abstract: Monitoring the well-being of infants is of utmost importance for ensuring their safety and providing timely care. This paper presents a novel approach to enhance traditional baby monitors by integrating intelligent classification of infant cries and real-time monitoring of sleep positions. The proposed smart baby monitor employs advanced signal processing techniques, machine learning algorithms, and computer vision methods to analyse audio and visual cues, providing valuable insights into a baby's needs and sleep patterns. This research paper presents the development of a Smart Baby Monitoring System equipped with one machine learning module and a wet diaper detection model. The system aims to enhance infant safety and improve caregiving practices. The first module employs YOLO and a Convolutional Neural Network (CNN) for position detection, identifying whether the baby is on their belly, back, or side. The position detection model achieves an average accuracy of 90%, with precision and recall rates 85%, ensuring reliable identification of the baby's needs Second module is a wet diaper detection module is incorporated to detect if the baby's diaper needs changing. The integration of these machine learning modules in the Smart Baby Monitoring System offers realtime monitoring, timely alerts, and improved infant care.

Intelligent recognition method of laser image based on big data analysis technology

In order to improve the recognition effect of laser images, this study designed an intelligent recognition method of laser images based on big data analysis technology. On the basis of setting up the laser holographic scanning device and parameters, the laser image is obtained by using the calibration method of vision system. In order to avoid the limitation of coordinate system in the process of laser image recognition, a rational function model with general attributes is constructed. Then, convolutional neural network is used to output the feature data of laser images, and Spark parallel support vector machine algorithm is used to complete the classification of laser images. Finally, the SVM classification model based on the big data analysis technology is constructed. The texture feature data can be input to quickly output the classification results of laser images, and then the intelligent classification and recognition of laser images can be realized according to the probability distribution. Experimental results show that this method can accurately identify the tiny features in laser images, and the recognition results have high peak signal-to-noise ratio and high recognition accuracy.

Intelligent classification of coal seams using spontaneous combustion susceptibility in IoT paradigm

ABSTRACT Coal is the cheapest source of energy. It is much cheaper than nuclear and petroleum. However, spontaneous combustion is inherent in coal seams that lead to fires at their place of storage. Therefore, early detection is required to ensure safety. This article aims to develop an intelligent system for the classification of coal seams based on their spontaneous heating on the Internet of Things (IoT) paradigm. It utilizes an artificial intelligence (AI) approach by utilizing a two-stage neural network architecture to render an accurate and robust categorization of coal samples. Three commonly available proximate analysis parameters of the coal samples have been used as input to this system. These inherent parameters are found to be sufficient to classify coal samples into different categories according to their combustion susceptibility. Three publicly available datasets are involved in the assessment purpose. Our proposed method has outperformed previous approaches. The coal seams classification results agree with the real field experiences.

Detection of High Impedance Faults in Power Lines using Empirical Mode Decomposition with Intelligent Classification Techniques

An uninterrupted and well-managed energy distribution system requires an effective protection system against various faults and uncertainties. Power system networks have a built-in issue called High Impedance Faults (HIFs), which can't be detected by standard protection mechanisms. This paper deals with the aforementioned issue by combining a deep learning technique (DLT) with a signal decomposition method known as Empirical Mode Decomposition (EMD). This approach utilizes the EMD for the extraction of multiple features from the input current signal, and LSTM neural network is employed for the accurate classification of HIF signals. The comparison results with various EMD methods and various switching events are also performed. Finally, this paper proposes an approximate location identification of HIFs. The simulation results demonstrate the effectiveness of the proposed method for detection in various fault scenarios.

Classifying Heart-Sound Signals Based on CNN Trained on MelSpectrum and Log-MelSpectrum Features.

The intelligent classification of heart-sound signals can assist clinicians in the rapid diagnosis of cardiovascular diseases. Mel-frequency cepstral coefficients (MelSpectrums) and log Mel-frequency cepstral coefficients (Log-MelSpectrums) based on a short-time Fourier transform (STFT) can represent the temporal and spectral structures of original heart-sound signals. Recently, various systems based on convolutional neural networks (CNNs) trained on the MelSpectrum and Log-MelSpectrum of segmental heart-sound frames that outperform systems using handcrafted features have been presented and classified heart-sound signals accurately. However, there is no a priori evidence of the best input representation for classifying heart sounds when using CNN models. Therefore, in this study, the MelSpectrum and Log-MelSpectrum features of heart-sound signals combined with a mathematical model of cardiac-sound acquisition were analysed theoretically. Both the experimental results and theoretical analysis demonstrated that the Log-MelSpectrum features can reduce the classification difference between domains and improve the performance of CNNs for heart-sound classification.

Research on an Integrated Intelligent Classification Algorithm Based on K-Means PCA-RF Machine Learning

With the rapid development of machine learning and artificial intelligence, research on classification models is gradually becoming popular. This article aims to propose a general classification model and classify indicator features. First, this paper constructs the data preprocessing based on K-Means, and data dimensionality reduction based on PCA algorithm. Finally, random forest algorithm (RF) is used for feature classification, and 325 groups of data are used for training. The results show that: (1) The K-Means PCA-RF algorithm constructed in this paper has good robustness and classification performance. (2) K-Means PCA-RF can effectively classify features and perform sensitivity analysis.

Іntelligent classification system based on ensemble methods

In the paper, based on machine learning methods, the solution of the classification task was investigated using a two-level structure of ensembles of models. To improve forecasting results, an ensemble approach was used: several basic models were trained to solve the same problem, with subsequent aggregation and improvement of the ob-tained results. The problem of classification was studied. The architecture of the intelli-gent classification system is proposed. The system consists of the following components: a subsystem of preprocessing and data analysis, a subsystem of data distribution, a subsystem of building basic models, a subsystem of building and evaluating ensembles of models. A two-level ensemble structure was used to find a compromise between bias and variance inherent in machine learning models. At the first level, an ensemble based on stacking is implemented using a logistic regression model as a metamodel. The pre-dictions that are generated by the underlying models are used as input for training in the first layer. The following basic models of the first layer were chosen: decision trees (DecisionTree), naive Bayesian classifier (NB), quadratic discriminant analysis (QDA), logistic regression (LR), support vector method (SVM), random forest model (RF). The bagging method based on the Bagged CART algorithm was used in the second layer. The algorithm creates N regression trees using M initial training sets and averages the re-sulting predictions. As the basic models of the second layer, the following were chosen: the first-level model (Stacking LR), the model of artificial neural networks (ANN); the linear discriminant analysis (LDA) model and the nearest neighbor (KNN) model. A study of basic classification models and ensemble models based on stacking and bag-ging, as well as metrics for evaluating the effectiveness of the use of basic classifiers and models of the first and second level, was conducted. The following parameters were de-termined for all the methods in the work: prediction accuracy and error rate, Kappa statistic, sensitivity and specificity, accuracy and completeness, F-measure and area under the ROC curve. The advantages and effectiveness of the ensemble of models in comparison with each basic model are determined.

A survey: Breast Cancer Classification by Using Machine Learning Techniques

Breast cancer in general is a common and deadly disease. Early detection can significantly reduce the chances of death. Using automated feature extraction and classification algorithms, physicians' experience in diagnosing and detecting breast cancer can be aided. This paper focuses on various statistical and machine learning studies of mammography datasets for enhancing the accuracy of breast cancer diagnosis and classification based on various variables. The Naïve Bayes, the K-nearest neighbors (KNN), the Support Vector Machine (SVM), the Random Forest, the Logistic Regression, Multilayer Perceptron (MLP), fuzzy classifier, and Convolutional Neural Network (CNN) classifiers, are the most widely used technologies in this field. In this study, we provide an overview of the existing CAD systems based on artificial intelligence classification techniques and many types of medical image modalities. Potential research initiatives to build more efficient and accurate CAD systems have been investigated.

Fault Diagnostics and Faulty Pattern Analysis of High-Speed Roller Bearings Using Deep Convolutional Neural Network

Abstract In this article, vibration-based fault diagnostics and response classification have been done for defective high-speed cylindrical bearing operating under unbalance rotor conditions. An experimental study has been performed to capture the vibration signature of faulty bearings in the time domain and for different speeds of the unbalanced rotor. Two-dimensional phase trajectories are generated by estimating the time delay and embedding dimension corresponding to vibration signatures. Qualitative analysis involves the implementation of a deep convolutional neural network (DCNN) utilizing the phase portraits as input to classify the nonlinear vibration responses. Comparison with the state-of-art classifiers such as artificial neural network (ANN), deep neural network (DNN), and k-nearest neighbor (KNN) is presented based on classification accuracy values. Thus, the values obtained are 61%, 67%, 72%, and 99% for ANN, DNN, KNN, and DCNN, respectively. Hence, the proposed intelligent classification model accurately identifies the dynamic behavior of bearing under unbalanced rotor conditions.

A Review on Quality Scanner Device for Radish and Carrot using Image Processing Technique

Abstract: Agriculture is the backbone of the Indian economy and productivity. The market value obtained from them is immense and effect both the export and quality valuation . Although grading and quality can be analysed by humans it is still inaccurate, time intensive, unreliable, subjective, and also hard which can be easily influenced by surrounding. Hence, there is an ultimate need for intelligent classification for vegetables. This paper focuses on Developing an analysis device for Radish (Raphanus raphanistrum ) and Carrot (Daucus carota subsp. sativus) which is an edible root vegetable grown in most parts of India. Radish and carrots are good sources of natural nitrates, minerals, calcium, and potassium. Thus, vegetables must be used as soon as possible to get a superior fresh character from them. External parameters such as size, shape, and quality factors of radish are important for finding out the total quality that is acceptable to the people. Various methods and algorithms are been put forwarded for the quality evaluation for the external defects present in it. Convolutional Neural Networks is the method used in image processing as this is more advantageous since it can learn special characteristics from images and by establish CNN for image classification where analyst have to gather the dataset. The requirement for doing this project is to have a computer and a python program for use the driver code. Image processing method is used for estimating quality of the particular food product efficiently in these days. This review paper examines current developments in methods for assessing food quality through image processing. The use of CNN (Convolutional Neural Networks) to identify characteristics of vegetables and its diseases through several built-in python libraries

Standardized Database of 12-Lead Electrocardiograms with a Common Standard for the Promotion of Cardiovascular Research: KURIAS-ECG.

Electrocardiography (ECG)-based diagnosis by experts cannot maintain uniform quality because individual differences may occur. Previous public databases can be used for clinical studies, but there is no common standard that would allow databases to be combined. For this reason, it is difficult to conduct research that derives results by combining databases. Recent commercial ECG machines offer diagnoses similar to those of a physician. Therefore, the purpose of this study was to construct a standardized ECG database using computerized diagnoses. The constructed database was standardized using Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) and Observational Medical Outcomes Partnership-common data model (OMOP-CDM), and data were then categorized into 10 groups based on the Minnesota classification. In addition, to extract high-quality waveforms, poor-quality ECGs were removed, and database bias was minimized by extracting at least 2,000 cases for each group. To check database quality, the difference in baseline displacement according to whether poor ECGs were removed was analyzed, and the usefulness of the database was verified with seven classification models using waveforms. The standardized KURIAS-ECG database consists of high-quality ECGs from 13,862 patients, with about 20,000 data points, making it possible to obtain more than 2,000 for each Minnesota classification. An artificial intelligence classification model using the data extracted through SNOMED-CT showed an average accuracy of 88.03%. The KURIAS-ECG database contains standardized ECG data extracted from various machines. The proposed protocol should promote cardiovascular disease research using big data and artificial intelligence.

A Comparative Study of Discrete Direction vs. Continuous Distance-Based Cost Function in Energy Classification

Over the past decades, different classification approaches with different characteristics have been developed to achieve more efficient and accurate results. Although the loss function used in the training procedure is a significant influential factor in the performance of classification models, it has been less considered. In general, in previous research, two main categories of continuous and semi-continuous distance-based loss functions are often applied to estimate the unknown parameters of classification models. Among these, continuous distance-based cost functions are among the most commonly used and most popular loss functions in diverse statistical and intelligent classifiers. In particular, the fundamental of this category of the loss functions is based on the continuous reduction of the distance between the fitted and actual values with the aim of improving the performance of the model. However, since the goal function of classification models belongs to the class of discrete ones, the application of learning procedures based on a continuous distance-based function is not coordinated with the nature of these problems. Consequently, it is theoretically illogical and practically at least inefficient. Accordingly, in order to fill this research gap, the discrete direction-based loss function in the form of mixed-integer programming is proposed in the training procedure of statistical, shallow/deep intelligent classifiers. In this paper, the impact of the loss function type on the classification rate of the classifiers in the energy domain is investigated. For this purpose, the logistic regression (LR), multilayer perceptron (MLP), and deep multilayer perceptron (DMLP), which are respectively among the most widely used statistical, shallow intelligent, and deep learning classifiers, are exemplarily chosen. Numerical outcomes from 13 benchmark energy datasets show that, in all benchmarks, the performances of the discrete direction learning-based classifiers, i.e., discrete learning-based logistic regression (DILR), discrete learning-based multilayer perceptron (DIMLP), and discrete learning-based deep multilayer perceptron (DIDMLP), is higher than its conventional versions. In addition, the proposed DILR, DIMLP, and DIDMLP models can on average yield an 89.88%, 94.53%, and 96.02% classification rate, which indicate a 6.78%, 5.90%, and 4.69% improvement from the classic versions, which only produce an 84.17%, 89.26%, and 91.72% classification rate. Consequently, the discrete direction-based learning methodology can be a more suitable, effective, and valuable alternative for training processes in statistical and shallow/deep intelligent classification models.

Knowledge-Based Features for Speech Analysis and Classification: Pronunciation Diagnoses

Accurate pronunciation of speech sounds is essential in communication. As children learn their native language, they refine the movements necessary for intelligible speech. While there is variability in the order of acquisition of speech sounds, there are some sounds that are more complex and are later developing. The rhotic /r/ is a later-developing sound in English, and some children require intervention to achieve accurate production. Additionally, individuals learning English as a second language may have difficulty learning accurate /r/ production, especially if their native language does not have an /r/, or the /r/ they produce is at a different place of articulation. The goal of this research is to provide a novel approach on how a knowledge-based intelligence program can provide immediate feedback on the accuracy of productions. In the proposed approach, the audio signals will first be detected, after which features of audio signals will be extracted, and finally, knowledge-based intelligent classification will be performed. Based on the obtained knowledge and application scenarios, novel features are proposed and used to classify various speaker scenarios.

Grid data asset relationship and intelligent classification integrating knowledge graph and Internet of Things

Abstract With the development of smart grids, power grids have accumulated massive amounts of data in various links such as power generation, transmission, substation, distribution, power consumption, and dispatch. More and more big data applications are beginning to be applied in various professional fields of the power grid. Promote the application and value discovery of smart grid big data through data fusion inside and outside the grid. Grid data has become an important asset for enterprise development, but power grid enterprises lack effective technical means to solve the whole life cycle monitoring and relationship of power grid data assets. Aiming at the relationship between power grid data assets, this paper proposes a set of grid data asset relationship and intelligent classification framework that integrates knowledge graph and Internet of Things. First, the grid knowledge graph extraction relationship is carried out by ProjE algorithm. Then, the relationship between power grid data assets and intelligent classification framework that integrates knowledge graph and Internet is proposed. Finally, the corresponding classification application is proposed by using intelligent classification algorithm. Experimental results show that the intelligent classification accuracy rate can reach 93.12% under the relationship between the knowledge graph and the Internet data assets, which has a new idea for the future development of the relationship between power grid data assets.

Fetal brain activity and the free energy principle.

To study whether the free energy principle can explain fetal brain activity and the existence of fetal consciousness via a chaotic dimension derived using artificial intelligence. In this observational study, we used a four-dimensional ultrasound technique obtained to collect images of fetal faces from pregnancies at 27-37weeks of gestation, between February and December 2021. We developed an artificial intelligence classifier that recognizes fetal facial expressions, which are thought to relate to fetal brain activity. We then applied the classifier to video files of facial images to generate each expression category's probabilities. We calculated the chaotic dimensions from the probability lists, and we created and investigated the free energy principle's mathematical model that was assumed to be linked to the chaotic dimension. We used a Mann-Whitney test, linear regression test, and one-way analysis of variance for statistical analysis. The chaotic dimension revealed that the fetus haddense and sparse states of brain activity, which fluctuated at a statistically significant level. The chaotic dimension and free energy were larger in the sparse state than in the dense state. The fluctuating free energy suggests consciousness seemed to exist in the fetus after 27weeks.