Quantizes real data and applies discrete bayes rule to achieve highest expected value.
Uses bruteforce hill climbing to find optimal boundaries for quantization.
Uses K smoothing to smooth probabilities of low probability intervals.
See write up for results.
#Setup: pip install -r requirements.txt
download observations.db from https://drive.google.com/open?id=1x-JfYVIih0mbSA1_D1Li1ZTf486l0zdT and place it in the root of this project.
#Run Tests: python -m tests.test_discrete_bayes python -m tests.test_quantizer
#Run Optimizations: to find optimal boundaries: python -m unittest tests.test_optimizer.TestOptimizer.test_optimize_bounds
to evaluate bounds: python -m unittest tests.test_optimizer.TestOptimizer.test_eval_bounds