A Comparative Study of Diabetes Detection Using The Pima Indian Diabetes Database

The accurate detection of diabetes plays a critical role in early intervention and effective management of the disease. In recent years, deep learning models have shown great potential in medical diagnosis tasks, including diabetes detection. This paper presents a comparative study of three popular models - Long Short-Term Memory (LSTM), Random Forest (RF), and Convolutional Neural Network (CNN) - for diabetes detection on the widely used Pima Indians Diabetes Database. The study aims to evaluate the performance of these models using common evaluation metrics, such as accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC). The dataset is preprocessed to handle missing values, normalize features, and split into training and testing sets. Each model is trained on the training set and evaluated on the testing set. The results of the study show that the LSTM model achieves the highest performance across all metrics. It demonstrates the ability to capture the temporal nature of the dataset and extract meaningful patterns for accurate diabetes detection. RF and CNN models also exhibit promising performance but slightly lower metrics compared to LSTM. In the comparative analysis, the strengths and weaknesses of each model are discussed. LSTM, as a recurrent neural network, excels in capturing temporal dependencies, while RF offers simplicity and interpretability. CNN, although originally designed for image analysis, shows potential when adapted to tabular data. The findings of this study have implications for healthcare practitioners and researchers working on diabetes detection. The LSTM model achieves its highest accuracy at 85%, demonstrating its effectiveness as an accurate method for predicting diabetes using the Pima Indians Diabetes Database (PIDD). However, it is important to acknowledge the limitations of the study, such as the relatively small dataset size and potential class imbalance in the dataset. Future research can address these limitations and further investigate the application of deep learning models in diabetes detection