Detecting whether or not a cell is cancerous is of vital importance to medical practitioners. Many times the data gathered about the cells of a tumor are not easily human readable, and have complicated relationships. The question is, how can we use statistically data driven methods to determine if a particular sample cell is cancerous?
Solution:
Develop a supervised classification model to predict the likelihood of a given cell from a tumor to be malignant or benign.
Methods:
Custom random forest implementation, and a TAN Bayesian network.
Frameworks and Platforms:
Python, custom modeling code
Outcomes:
Developed a cancer classification system with over 90% accuracy in classifying cells as malignant or not.
