The AUTOMATED BREAST CANCER PREDICTION WITH EXPONENTIAL LINEAR UNIT-ACTIVATED BILSTM AND ATTENTION-BASED MECHANISMS USING HISTOLOGICAL AND LYMPH NODE DATA

BREAST CANCER PREDICTION

Abstract

Background Information: Because of the serious health risks associated with breast cancer, early detection and an accurate prognosis are essential for effective therapy. Accurate evaluation of cellular properties and molecular subtypes is enhanced by sophisticated diagnostic models.

Objective: To increase the precision of cellularity and subtype detection, an automated breast cancer prediction model utilizing Bi STM and attention-driven ELU activations is being developed.

Methods: The model uses BiLSTM networks with attention mechanisms and ELU activations to interpret data on lymph nodes and histology. Robotic automation improves the consistency and efficiency of the system's operations.

Results: The suggested model outperformed earlier diagnostic models with scores of 98.65% accuracy, 98.01% precision, 99.12% recall, and 98.98% F1-score.

Conclusion: By precisely predicting breast cancer risk and identifying molecular subtypes, this sophisticated model provides a potent instrument that enhances early detection and prognosis.