Using machine learning in breast cancer diagnosis through ultrasonography

Abstract

Objective: To apply machine learning techniques to evaluate the efficiency and accuracy in diagnosing breast cancer using ultrasound images. Methods: Using the Orange Data Mining tool, images from a dataset obtained on the Kaggle platform containing ultrasound scans of breasts with benign and malignant nodules were processed. Results: Image analysis in the tests was performed using machine learning techniques, Neural Network and KNN, which demonstrated high accuracy rates in classifying benign and malignant nodules with results above 90% accuracy. Conclusion: The research emphasizes the importance of machine learning tools to create robust and accurate predictive models, significantly improving clinical diagnostic accuracy and increasing patient recovery chances.