Leveraging Intelligent Tools and Techniques for Early Breast Cancer Detection Using Demographic Data

Detecting breast cancer early is extremely important for a successful treatment. Current day clinical diagnosis methods are extremely expensive and time-consuming. This paper aims to change that hence making diagnosis cheap, fast and reliable. A dataset using demographic data only was used. The dataset underwent preprocessing to eliminate unwanted data. An AI model was made using Random Forest and gradient boosting algorithms. To tackle the Imbalance in the data set a combination of SMOTE and random under-sampling was used. This AI model showed an average accuracy of 89.6% and an AUC-ROC score of 0.78. These promising results show the viability of this model. Via a simple survey, a cheap accurate and rapid breast cancer detection system can be developed. The viability of using purely demographic data is tested and proven. However, there is still room for improvement in the model. The difference between false positives and true positives needs to be reduced before any practical use.