ABSTRACT
In a period of constantly increasing healthcare technology, reliable and accurate methods for diagnosing and predicting life-threatening disorders, including liver cancer, are more important than ever. In this work, we study machine learning and explore potential applications of Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), k-Nearest Neighbours (KNN), Support Vector Machines (SVM), and CNNs to improve the diagnosis of liver cancer. A substantial dataset of 1320 medical images from CT and MRI scans supports our research and enables an in-depth examination of state-of-the-art methods for illness detection. We meticulously trained and evaluated every machine learning model using this dataset in order to ascertain its diagnostic proficiency. CNNs had the highest accuracy rate, at 96.7%; the results were encouraging and instructive. RNNs showed a great level of performance in interpreting sequential medical data. KNN and SVM demonstrated consistent accuracy, which increased its adaptability to a range of healthcare situations. This finding represents a major step forward in improving patient outcomes by facilitating early detection and possibly changing the diagnosis of liver cancer. By integrating machine learning into clinical practise and facilitating more accurate, efficient, and prompt detection and treatment for liver cancer, we think this innovative strategy will significantly progress medical science.
