AZUA, SUURSHATER STEPHEN (2026) MammoCareAI:Development of Breast Cancer Detection and Risk Prediction System Using Convolutional Neural Networks. Other thesis, Godfrey Okoye University, Enugu Nigeria.
|
Text
Formatted Development of Breast Cancer Detection and Risk Prediction.pdf Download (1MB) | Preview |
Abstract
Cancer of the breast has contributed to most cancer deaths among women all over the world, which is why there is a significant emphasis on early detection of breast cancer and on the ability to accurately assess a woman's risk of breast cancer. In this study, we sought to create a breast cancer detection and risk prediction system using CNNs to assist in the early diagnosis of breast cancer and in the development of a clinical decision regarding a patient's treatment. The initial phase of the study was to identify the major risk factors linked to the development of breast cancer, followed by the gathering of enough quality dataset, which comprises both benign and malignant breast cancer cases. This dataset served as the foundation for analysing the relationship between identified risk factors and breast cancer occurrence. The data preparation phase included data cleaning, normalising, and transforming the data into a format appropriate for training the CNN model. Next, we created the CNN model that could learn the discriminative characteristics of the benign and malignant breast cancer cases and trained it to detect the presence of these characteristics. After the model was trained, we created a web-based user interface to ease the identification of risk and the practical use of the predictive model. Lastly, we evaluated the performance of the predictive model using multiple metrics such as accuracy, precision, recall, F1-score, specificity, confusion matrix, and Receiver Operator Characteristic (ROC) curve. The results obtained after model evaluation demonstrate that the proposed model performed exceptionally well in breast cancer classification, as it achieved an accuracy of 94.12%, precision of 93.10%, recall of 95.29%, F1-score of 94.19%, specificity of 92.94%, and an Area Under the Curve (AUC) value of 96.87%. The results confirm the model's strong ability to accurately distinguish between different breast cancer classes, which makes it a reliable tool for prediction and classification tasks. The developed system provides an effective and reliable tool, which will be employed in order to carry out breast cancer detection and risk prediction accurately, as well as the potential to support healthcare professionals in enhancing early diagnosis and improving patient outcomes.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | Q Science > Q Science (General) |
| Divisions: | Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science |
| Depositing User: | Cynthia Ugwuoti |
| Date Deposited: | 22 Jul 2026 14:13 |
| Last Modified: | 22 Jul 2026 14:13 |
| URI: | http://eprints.gouni.edu.ng/id/eprint/6035 |
Actions (login required)
 |
View Item |
