Code repository for "Tumor Characterization using Unsupervised Learning of Mathematical Relations within Breast Cancer Data" by Cristian Axenie and Daria Kurz submitted at ENNS ICANN2020 (15th – 18th September 2020).

Codebase:

datasets - the experimental datasets (csv files) and their source, each in separate directories models - codebase to run and reproduce the experiments

Directory structure:

model/.

• create_init_network.m - init networks (SOM + HL)
• error_std.m - error std calculation function
• tumor_estimator_core.m - main script to run the system
• model_rmse.m - RMSE calculation function
• model_sse.m - SSE calculation function
• parametrize_learning_law.m - function to parametrize learning
• present_tuning_curves.m - function to visualize SOM tuning curves
• randnum_gen.m - weight initialization function
• tumor_growth_model_fit.m - function implementing ODE models
• tumor_growth_models_eval.m - main evaluation on runtime
• visualize_results.m - visualize output and internals
• visualize_runtime.m - visualize runtime

Usage:

• model/tumor_estimator_core.m - main function that runs the system and generates the runtime output file (mat file)
• model/tumor_growth_models_eval.m - evaluation and plotting function reading the runtime output file