Liver fibrosis, characterized by the excessive accumulation of extracellular matrix proteins, represents a significant global health concern with the potential to progress to cirrhosis and hepatocellular carcinoma if left untreated. Early and accurate diagnosis of liver fibrosis is critical for effective management and treatment, yet traditional diagnostic methods, such as liver biopsy, are invasive, costly, and carry inherent risks. This necessitates the development of non-invasive, reliable, and efficient diagnostic tools. This study explores the application of machine learning techniques for classifying liver fibrosis using the Indian Liver Patient Dataset (ILPD). The dataset underwent rigorous preprocessing, including imputation, transformation, and resampling, ensuring robust model training. Several algorithms, including Random Forest Gradient Boosting, XGBoost, Bagging classifier, ExtraTrees Classifier, Stacking Classifier, KNeighbors Classifier, Artificial Neural Networks (ANN), and a 1D Convolutional Neural Network (1D-CNN) were evaluated. The 1D-CNN achieved the highest accuracy of 97.14%, followed by the Stacking Classifier with 96.64%. SHAP (SHapley Additive exPlanations) has been used to interpret the 1D-CNN, providing insights into feature importance. The findings demonstrate the potential of machine learning for non-invasive, efficient, and reliable liver fibrosis diagnosis.