Diagnosis of polycystic ovary syndrome (PCOS) using TabNet and BiLSTM with attention model

Polycystic ovary syndrome (PCOS) is a common endocrine disorder that affects women of reproductive age and often leads to complications such as infertility, metabolic disorders, and hormonal imbalance. Early and accurate diagnosis of PCOS is crucial for effective treatment and control. In this study, we propose a novel hybrid deep learning model that integrates TabNet and BiLSTM with an attention mechanism for PCOS detection. The proposed model effectively captures both tabular data dependencies and sequential patterns and achieves an accuracy of 93.45%. To verify its effectiveness, we compare our model with several traditional machine learning and deep learning approaches, including Random Forest, XGBoost, CatBoost, CNN, RNN, and BERT. The experimental results show that our model outperforms these baselines in terms of accuracy, precision, recall, and F1-score. Integrating the interpretability of TabNet with the sequential learning capability of BiLSTM improves the representation of features.