article · Journal of King Saud University - Computer and Information Sciences
Accurate cancer prognosis is challenging because multiple factors contribute to the disease, meaning clinical assessment alone may be insufficient for classification. Deep neural learning provides a way to analyse substantial volumes of clinical and genetic information swiftly. A deep neural learning cancer prediction model, termed DNLC, uses a deep network to identify optimal features before training a deep neural network on genomic or clinical data samples. The model was evaluated across five distinct cancer datasets covering colon cancer, lung adenocarcinoma, squamous cell carcinoma, breast cancer, and leukaemia. Using an eighty-twenty split for training and testing data, the model achieved an average prediction accuracy of 93 per cent. This performance exceeded the accuracy levels attained by earlier convolutional neural network and recurrent neural network models, demonstrating an effective approach for predicting cancer in its earlier stages.
Cancer classification relies on complex combinations of genetic and clinical indicators that can challenge conventional diagnostic assessments. Automated assistance tools that reliably interpret these large datasets can support earlier disease identification. Demonstrating high predictive accuracy across diverse cancer types suggests that deep neural learning models can help improve the reliability of prognostic evaluations, assisting healthcare professionals in making timely clinical decisions.
This model presents potential applications as diagnostic support software for clinical oncologists and healthcare providers handling genomic and clinical cancer datasets. Given that the findings are based on experimental validation across five benchmark datasets using an eighty-twenty data split, the technology represents early-stage to applied algorithmic research. Real-world commercial deployment would require integration into clinical workflows, validation in prospective trials, and assessment against diverse patient cohorts in hospital environments.
AI-generated from the published abstract. Always read the original work before citing.
We know today that numerous factors play a significant role as causes of cancer. Because of this, a doctor's opinion alone cannot be used to classify cancer. Intelligent algorithms providing medical assistance are therefore necessary. In addition, many researchers have adopted them for estimating the likelihood of patient survival, and others have employed predictive methodologies like machine learning and deep learning to forecast prognoses for cancer. The accuracy of predictive cancer prognosis is currently of widespread concern. Since deep neural learning (DNL) methods can quickly predict outcomes from a significant amount of clinical and genetic data, they are essential for predicting various diseases. Deep neural learning is the foundation of our suggested approach. Our deep neural learning cancer prediction model (DNLC) has the following stages. In the first stage, Deep Network (DN) is used to select the best collection of features from datasets. In the second stage, we train genomic or clinical data samples with a deep neural network (DNN). In the third stage, we evaluate the capabilities of the DNLC model of predicting cancer in its earlier stages. For classification, DNLC uses five cancer datasets, which are for colon, lung adenocarcinoma, squamous cell carcinoma, breast, and leukaemia cancers. The five cancer datasets are used in experiments to predict how well the suggested model will perform. The dataset is divided into two parts: training sets, which make up 80% of the dataset, and testing sets, which make up 20%. The experimental results show that the suggested model performs better in terms of accuracy than earlier CNN and RNN models. Our findings demonstrate that the DNLC technique, with an average accuracy of 93%, outperforms other methods in all circumstances.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1016/j.jksuci.2023.101565
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
