An Investigation on Machine Learning Models in Classification and Identifications Cervical Cancer Using MRI Scan

This study analyzes the effectiveness of machine learning models in the classification of cervical cancer using a dataset of 900 cancer and 200 non-cancer images gathered from online resources and hospitals. The dataset, covering both CT and MRI images, undergoes rigorous preprocessing, including standardization, normalization, and noise reduction, to enhance its quality for model training. Four machine learning models, namely VGG16, CNN, KNN, and RNN, are recruited to predict cancer and non-cancer cases. During the testing phase, VGG16 emerges as the most accurate, achieving an impressive accuracy of 95.44%, followed by CNN at 92.3%, KNN at 89.99%, and RNN at 86.233%. Performance parameters, such as precision, recall, F1 score, and accuracy, are fully analyzed, providing insights into each model's strengths and capabilities. These discoveries not only contribute to the advancement of cervical cancer diagnostic techniques but also underscore the potential of machine learning in medical imaging. The study emphasizes the relevance of model selection and provides a framework for future research endeavors seeking to enhance the accuracy and performance of cervical cancer diagnosis through the merger of advanced computational techniques with standard diagnostic practices.