Publications Scientifiques

Permanent URI for this communityhttps://dspace.univ-boumerdes.dz/handle/123456789/10

Browse

Has files: YesSubject: search.filters.subject.Classification

Search Results

Now showing 1 - 10 of 17
  • No Thumbnail Available
    Item
    New criteria for wrapper feature selection to enhance bearing fault classification
    (SAGE, 2023) Sahraoui, Mohammed Amine; Rahmoune, Chemseddine; Meddour, Ikhlas; Bettahar, Toufik; Zair, Mohamed
    Classification is a critical task in many fields, including signal processing and data analysis. The accuracy and stability of classification results can be improved by selecting the most relevant features from the data. In this paper, a new criterion for feature selection using wrapper method is proposed, which is based on the evaluation of the classification results according to the accuracy and stability (standard deviation) of each class and the number of selected features. The pro- posed method is evaluated using Random Forest (RF) and Ant Colony Optimization (ACO) algorithms on a benchmark dataset. Results show that the proposed method outperforms classical feature selection methods in terms of accuracy and stability of classification results, especially for the difficult-to-classify combined damage class. This study demon- strates the effectiveness of the proposed new wrapper feature selection criterion to improve the performance of classifi- cation algorithms with higher stability (STD: C1 = 0.5, C2 = 0.8, C3 = 0.6, C4 = 1.8) and better accuracy (average C1 = 98.5%, C2 = 96.6%, C3 = 9.5%, C4 = 93) for the both; the statoric current and the vibration signal compared to other techniques. Machine learning methods had proven their efficiency in time-varying machines fault diagnosis when taking vibration signals and statoric currents extracted features as inputs. However, the use of the both demonstrated a higher robustness and a remarkable superiority.
  • No Thumbnail Available
    Item
    Local feature selection using the wrapper approach for facial-expression recognition
    (Scientific Research Support Fund of Jordan (SRSF) and Princess Sumaya University for Technology (PSUT), 2024) Boukhobza, Fatima Zohra; Gharbi, Abdenour Hacine; Rouabah, Khaled; Ravier, Phillipe
    Automatic Facial Expression Recognition (FER) systems provide an important way to express and interpret the emotional and mental states of human beings. These FER systems transform the facial image into a set of features to train a classifier capable of distinguishing between different classes of emotions. However, the problem often posed is that the extracted feature vectors sometimes contain irrelevant or redundant features, which decreases the accuracy of the induced classifier and increases the computation time. To overcome this problem, dimensionality must be reduced by selecting only the most relevant features. In this paper, we study the impact of adding the "Wrapper" selection approach and using the information provided by different local regions of the face such as the mouth, eyes and eyebrows, on the performance of a traditional FER system based on a local geometric feature-extraction method. The objective here is to test and analyze how this combination can improve the overall performance of the original traditional system. The obtained results, based on the Multimedia Understanding Group (MUG) database, showed that the FER system combined with the proposed feature-selection strategy gives better classification results than the original system for all four classification models; namely, K-Nearest Neighbor (KNN) classifier, Tree classifier, NB classifier and Linear Discriminant Analysis (LDA). Indeed, a considerable reduction (up to 50%) in the number of features used and an accuracy of 100%, using the LDA classifier, were observed, which represents a significant improvement in terms of computation time, efficiency and memory space. Furthermore, the majority of relevant features used are part of the "eyebrows’ region", which proves the importance of using information from local regions of the face in emotion recognition tasks.
  • No Thumbnail Available
    Item
    Integration of generalized discriminant analysis and classification technique for identification well test interpretation model
    (IJISAE, 2023) Saifi, Redha; Zeraibi, Nourreddine
    This paper presents a hybrid method that combines generalized discriminant analysis and machine learning technique for identifying well test model. The proposed method consists of three stages: (1) nonlinear combination of features spaces to maximize the separability among the class models through generalized discriminant analysis. (2) Construction a set of classifier and classify the new data points by a plurality vote of their prediction. This method is described in detail to ensure successful replication of results. The required training and test dataset were generated by using analytical solution models. In our case, there were used 600 samples: 70% for training, 15% for validation, and 15% for testing. We notice that the generalized discriminant analysis is an effective strategy in reducing the number of input features, simplifying the network structure, and lowering the training time of the ANN. The results obtained show that the proposed model provides better performance when predicting new data with a coefficient of correlation approximately equal to 99%.
  • No Thumbnail Available
    Item
    Machine learning-based Shapley additive explanations approach for corroded pipeline failure mode identification
    (Elsevier Ltd, 2024) Ben Seghier, Mohamed El Amine; Mohamed, Osama Ahmed; Ouaer, Hocine
    Rapid failure mode identification of oil and gas pipelines can prevent catastrophic consequences, improve fast intervention and enhance the design safety of these critical systems. This paper proposes explainable-based machine learning models using to determine the failure mode of corroded pipelines as a function of geometric configurations, material properties, and corrosion defect details. To determine the best identification model, this study examined eight machine learning models, including Nave Bayes, K-Nearest Neighbors, Decision Tree, Random Forest, Adaptive Boosting, Extreme Gradient Boosting, Light Gradient Boosting Machine, and Category Boosting, based on a comprehensive experimental database for steel pipelines with various corrosion/crack defect configurations. Furthermore, the Shapley additive explanations approach is utilized to rank the input variables for failure mode identification and explains the machine learning model predicting a specific failure mode for a given sample. In identifying the failure mode of corroded pipelines, the proposed Extreme Gradient Boosting model indicated the highest accuracy in term of performance evaluation compared to all other proposed models. In addition, the model-explanation findings show that the important parameter influencing the failure mechanism of corroded pipelines is the depth of corrosion defects followed by the pipeline wall thickness. The proposed framework is adaptable enough to allow further use of experimental results for having new insights.
  • No Thumbnail Available
    Item
    EEG signal feature extraction and classification for epilepsy detection
    (Slovene Society Informatika, 2022) Cherifi, Dalila; Falkoun, Noussaiba; Ouakouak, Ferial; Boubchir, Larbi; Nait-Ali, Amine
    Epilepsy is a neurological disorder of the central nervous system, characterized by sudden seizures caused by abnormal electrical discharges in the brain. Electroencephalogram (EEG) is the most common technique used for Epilepsy diagnosis. Generally, it is done by the manual inspection of the EEG recordings of active seizure periods (ictal). Several techniques have been proposed throughout the years to automate this process. In this study, we have developed three different approaches to extract features from the filtered EEG signals. The first approach was to extract eight statistical features directly from the time-domain signal. In the second approach, we have used only the frequency domain information by applying the Discrete Cosine Transform (DCT) to the EEG signals then extracting two statistical features from the lower coefficients. In the last approach, we have used a tool that combines both time and frequency domain information, which is the Discrete Wavelet Transform (DWT). Six different wavelet families have been tested with their different orders resulting in 37 wavelets. The first three decomposition levels were tested with every wavelet. Instead of feeding the coefficients directly to the classifier, we summarized them in 16 statistical features. The extracted features are then fed to three different classifiers k-Nearest Neighbor (k-NN), Support Vector Machine (SVM), and Artificial Neural Network (ANN) to perform two binary classification scenarios: healthy versus epileptic (mainly from interictal activity), and seizure-free versus ictal. We have used a benchmark database, the Bonn database, which consists of five different sets. In the first scenario, we have taken six different combinations of the available data. While in the second scenario, we have taken five combinations. For Epilepsy detection (healthy vs epileptic), the first approach performed badly. Using the DCT improved the results, but the best accuracies were obtained with the DWT-based approach. For seizure detection, the three methods performed quite well. However, the third method had the best performance and was better than many state-of-the-art methods in terms of accuracy. After carrying out the experiments on the whole EEG signal, we separated the five rhythms and applied the DWT on them with the Daubechies7 (db7) wavelet for feature extraction. We have observed that close accuracies to those recorded before can be achieved with only the Delta rhythm in the first scenario (Epilepsy detection) and the Beta rhythm in the second scenario (seizure detection)
  • No Thumbnail Available
    Item
    Deep Learning Models for Intracranial Hemorrhage Recognition: A comparative study
    (Elsevier, 2022) Ammar, Mohammed; Lamri, Mohamed Amine; Mahmoud, Saïd; Laid, Amel
    Every day, a large number of people with brain injury are received in the emergency rooms. Due to the large number of slices analyzed by the doctors for each patient and to accelerate the diagnosis, the development of a precise computer-aided diagnosis system becomes very recommended. The aim of our work is developing a tool to help radiologists in the detection of intracranial hemorrhage (ICH) and its five (05) subtypes in computed tomography (CT) images. Five deep learning models are tested: ResNet50, VGG16, Xception, InceptionV3 and InceptionResNetV2. Before training these models, preprocessing operations are performed like normalization and windowing. The experiments show that VGG-16 architecture provides the best performances. The model achieves an accuracy of 96%.
  • No Thumbnail Available
    Item
    Gearboxes fault detection under operation varying condition based on MODWPT, Ant colony optimization algorithm and Random Forest classifier
    (SAGE Publications, 2021) Ikhlef, Boualem; Rahmoune, Chemseddine; Bettahar, Toufik; Benazzouz, Djamel
    Gearboxes are massively utilized in nowadays industries due to their huge importance in power transmission; hence, their defects can heavily affect the machines performance. Therefore, many researchers are working on gearboxes fault detection and classification. However, most of the works are carried out under constant speed conditions, while gears usually operate under varying speed and torque conditions, making the task more challenging. In this paper, we propose a new method for gearboxes condition monitoring that is efficiently able to reveal the fault from the vibration signatures under varying operating condition. First, the vibration signal is processed with the Maximal Overlap Discrete Wavelet Packet Transform (MODWPT) to extract the AM-FM modes. Next, time domain features are calculated from each mode. Then the features set are reduced using the Ant colony optimization algorithm (ACO) by removing the redundant and unimportant parameters that may mislead the classification. Finally, an ensemble learning algorithm Random Forest (RF) is used to train a model able to classify the fault based on the selected features. The innovative aspect about this method is that, unlike other existing methods, ACO is able to optimize not only the features but also the parameters of the classifier in order to obtain the highest classification accuracy. The proposed method was tested on varying operating condition real dataset consisting of six different gearboxes. In the aim to prove the performance of our method, it had been compared to other conventional methods. The obtained results indicate its robustness, and its accuracy stability to handle the varying operating condition issue in gearboxes fault detection and classification with high efficiency
  • No Thumbnail Available
    Item
    Detection of power quality disturbances using discrete wavelet transform
    (IEEE, 2017) Dekhandji, Fatma Zohra
    With the growing use of sensitive and susceptive electronic and computing equipment, power quality is foreseen to cause a great concern to electric utilities. The best analysis on power quality is vital to provide better service to customers. Disturbances in power system usually produce continuity changes in the power signal. Wavelet transform is particularly useful in detecting discontinuities in signals, and this makes it appropriate for detection of disturbances in power quality. Wavelet transform is proposed to detect and identify the power quality disturbance at its instance of occurrence. Power quality disturbances are sag, swell, interruption, transient and harmonic. This study reviews various kinds of power quality disturbances with the goal of detecting them using wavelet transform. The results show clearly various forms of changes in amplitude and frequency of the signals. The application shows that this method is fast, sensitive, and practical for detection and identification of power quality disturbance
  • No Thumbnail Available
    Item
    Intelligent detection without modeling of behavior unusual by fuzzy logic
    (Springer, 2017) Chebi, Hocine; Acheli, Dalila; Kesraoui, Mohamed
  • No Thumbnail Available
    Item
    Signal processing deployment in power quality disturbance detection and classification
    (2017) Dekhandji, Fatma Zohra
    Power quality disturbances have adverse impacts on the electric power supply as well as on the customer equipment. Therefore, the detection and classification of such problems is necessary. In this paper, a fast detection algorithm for power quality disturbances is presented. The proposed method is a hybrid of two algorithms, abc–0dq transformation and 90 phase shift algorithms. The proposed algorithm is fast and reliable in detecting most voltage disturbances in power systems such as voltage sags, voltage swells, voltage unbalance, interrupts, harmonics, etc. The three-phase utility voltages are sensed separately by each of the algorithms. These algorithms are combined to explore their individual strengths for a better performance. When a disturbance occurs, both algorithms work together to recognize this distortion. This control method can be used for critical loads protection in case of utility voltage distortion. Simulation and analysis results obtained in this study illustrate high performance of the strategy in different single-phase and three-phase voltage distortions