HEPATITIS C VIRUS STAGES CLASSIFICATION USING TRANSIT SEARCH ALGORITHM AND LONG SHORT TERM MEMORY

Hepatitis C Virus (HCV) is a disease that infects the liver with multiple stages that spread through blood, requiring blood tests and body symptoms for diagnosis. The diagnosis should decide which stage the patient reaches. This work suggests a hybrid classification method based on optimization and machine learning techniques. A Long Short-Term Memory Neural Network (LSTM) is used as a classifier to identify the stages of the disease, based on the four stages, using 28 features comprising body symptoms and blood tests. To find the optimal number of hidden cells in the hidden layers of LSTM, the Transit Search Algorithm (TSA) is used, TSA identifies the best integer value for the number of hidden cells and selects the best features combined with this number of cells that will give the highest accuracy of classification. The preprocessing step is required to enhance the quality of the dataset and resizing it. In multiclass classification, a large dataset is essential for the network to learn effectively through all classes. To achieve this, Synthetic Minority Oversampling Techniques (SMOTE) is used to generate similar data that will increase the size of the dataset. The proposed TSA-LSTM method achieves high performance, with classification accuracy exceeding 99% outperforming previous works