Development of Machine Learning algorithms for Breast Cancer prediction


Tortós-Sala Granés, Carla  


Breast cancer, like cancer in general, is an uncontrolled growth of cells. That is, a malignant tumor that has developed in the cells of the breast. This tumor can begin to develop in the area of ​​the lobes, where the glands that secrete and produce milk are found, or in the duct, where the milk passes from the lobes to the nipple. However, there is also a small chance that this tumor originates in the stroma. One in eight women is diagnosed with breast cancer in her lifetime, making it the most common cancer among women. Despite this prevalence, when it is detected when it is in the initial stages, the probability of recovering is very high. The main objective of this TFG is to generate Machine Learning models to be able to predict the optimal treatment for cancer patients. We will work with K-NN (K-Nearest Neighbor) and Random Forest Classification models. In addition, we will carry out a study of transferability of the models generated between data sets. A feature importance study will also be carried out to identify how significant the variables are in the data set. The results of the study showed that, with only five parameters, the model could have a higher precision than that obtained with all the parameters, both validation and test. This demonstrates the feasibility of machine learning algorithms for accurate cancer prediction.



Fernández Esmerats, Joan 


IQS SE - Undergraduate Program in Pharmacy