Note: This code repository is no longer being maintained. Please refer to this repository instead for the more up-to-date (and more robust) version of the same work.

This repository contains the code used in the paper What is the value of Experimentation & Measurement?, which appeared in the IEEE ICDM 2019 conference.

Requirements: python>=3.6, numpy, scipy, pandas, matplotlib

Experiments, case studies, and empirical extensions in the paper can be found in the following notebooks:

• Section V - Empirical verification of the theoretical value of expectations and variances: src/theoretical_quantity_verification.ipynb
• Section VI - Case studies (e-commerce companies & marketing companies): src/value_gained_simulation.ipynb
• Appendix B-A - Empirical calculation of the risk: src/var_D_bound.ipynb
• Appendix B-B - Valuation Under Independent t-Distributed Assumptions: src/normal_t_comparison.ipynb
• Appendix B-C - Partial Estimation / Measurement Noise Reduction: src/partial_noise_reduction.ipynb