Yazar "Hazim, Layth Rafea" seçeneğine göre listele

Listeleniyor 1 - 3 / 3
  • Küçük Resim
    Öğe
    e-Diagnostic system for diabetes disease prediction on an IoMT environment-based hyper AdaBoost machine learning model
    (Springer, 2024) Jasim, Abdulrahman Ahmed; Hazim, Layth Rafea; Mohammedqasim, Hayder; Mohammedqasem, Roa’a; Ata, Oğuz; Salman, Omar Hussein
    One of the most fatal and serious diseases that humans have encountered is diabetes, an illness affecting thousands of individuals yearly. In this era of digital systems, diabetes prediction based on machine learning (ML) is gaining high momentum. One of the benefits of treating patients early in the course of their noncommunicable diseases (NCDs) is that they can avoid costly therapies when the illness worsens later in life. Incidentally, diabetes is complicated by the dearth of medical professionals in underserved areas, such as distant rural communities. In these situations, the Internet of Medical Things and machine learning (ML) models can be used to offer healthcare practitioners the necessary prediction tools to more effectively and timely make decisions, thus assisting the early identification and diagnosis of NCDs. In this study, four conventional and hyper-AdaBoost ML models were trained and tested on the PIMA Indian Diabetes dataset. Patients with diabetes were classified on the basis of laboratory findings. Pre-processing tasks, such as the handling of imbalanced data and missing values, were performed prior to feature importance and normalisation activities. The algorithm with the best performance was examined using precision, accuracy, F1, recall and area under the curve metrics. Then, all ML models were hyper parametrically tuned via grid search to optimise their performance and reduce their error times. The decision process was also evaluated to further enhance the models. The AdaBoost-ET model performed even when features were not selected for binary classification. The model proposed in this study can predict diabetes with unprecedented high accuracy compared with the models in previous studies.
  • [ X ]
    Öğe
    Intrusion Detection System (IDS) of multiclassification IoT by using pipelining and an efficient machine learning
    (Institute of Electrical and Electronics Engineers Inc., 2023) Hazim, Layth Rafea; Jasim, Abdulrahman Ahmed; Ata, Oğuz; Ilyas, Muhammad
    The Internet of Things (IoT) has quickly advanced and been incorporated into many different fields. With the use of IoT technology, gadgets can receive, process, and send data automatically. IoT has been rapidly accepted in many important fields since it makes life easier and increases service quality, yet it still faces significant privacy and security problems. An Intrusion Detection System (IDS) could be implemented as a security feature to protect IoT networks from a variety of cyberattacks. This study suggests using IDS to defend against a wide range of cyberattacks on IoT systems. The suggested approach makes use of the Multi-layer Perceptron (MLP) as well as Extra Trees (ExT) as efficient algorithms of classification. Also, the study uses the pipeline to put together several cross-validated phases while selecting various parameters to increase the detection rate. One dataset is utilized for evaluating and analyzing the performance outcomes so as to validate the efficiency of the suggested IDS approach. The evaluation findings show that the suggested IDS methods may greatly increase detection performance results concerning accuracy rate, precision, F1-score, and recall while also improving detection efficiency.
  • [ X ]
    Öğe
    Textual Authenticity in the AI Era: Evaluating BERT and RoBERTa with Logistic Regression and Neural Networks for Text Classification
    (Institute of Electrical and Electronics Engineers Inc., 2024) Hazim, Layth Rafea; Ata, Oğuz
    AI-generated content impersonating human writing is an issue that has gained attention as AI spreads its wings. This particular study serves as a comparison between the existing Logistic Regression and Feedforward Neural Networks (FNNs) by employing sentence-BERT-appended models and RoBERTa in the determination of authenticity in a given text. A dataset that was balanced between human-written and AI-generated texts was utilized. Techniques like tokenization and normalization were first used, followed by feature extraction using transformer-based models. Cross-validation and confusion matrix analysis that used measures such as accuracy, precision, recall, F1 score, and ROC AUC were included to guarantee the models' robustness. The hybrid RoBERTa-FNN model that was deposed challengers was the most outstanding model in respect of precision and recall, and the highest accuracy (99.95%) was obtained as mentioned in the data. The improved performance serves as a proof of how effectively RoBERTa uses its embeddings to represent context on the fine-grained level required for this kind of text classification. This work is a stepping stone to the creation of strong AI text detection systems, besides our advancements to the knowledge of models and embedding performance with respect to text classification. The results lay emphasis on the selection of model configuration and the embedding technique, as they are the key factors in achieving the best results in practical applications. © 2024 IEEE.