EARLY PREDICTION OF CERVICAL CANCER RISK USING EXPLAINABLE DEEP NEURAL NETWORKS AND FEATURE ENGINEERING OPTIMIZATION

Abstract

Early detection of cervical cancer plays an essential role in saving lives as well as improving outcomes for patients. The current study proposes an explainable deep learning technique to expect the likelihood of cervical cancer. It builds a classifier that identifies people who are at probability of mounting cervical cancer based on their demographics, behavior, and medical history. Data preprocessing techniques alike normalization, handling missing values, and feature transformation can be operated to improve the condition of input data as well as the efficiency of the classifier. The deep learning classifier discovers patterns and associations between risk factors. Explainable AI techniques are incorporated to improve predictability and highlight the importance of features. From the results of the experiments, it can be noted that the proposed framework operates appropriately in terms of accuracy, precision, recall, and F1-score. Thus, it can be recommended that the presentation of explainable deep learning is excellent for decision support in cervical cancer risk assessment.