dataset for Repetitive Readout Enhanced by Machine Learning

For the paper titled "Repetitive Readout Enhanced by Machine Learning"<br>The folder name corresponds to simulated NV photodynamic parameters <br> training.mat includes data_plus (5000*8000), data_minus (2500*8000), data_zero (2500*8000), corresponding to 14N starting in +1 (dark state), -1 (bright state), 0 (bright state) state (+1 is the dark state) <br> test.mat includes data_test (4000*8000), the first 1000 is |mI=-1&gt;, second 1000 |mI=0&gt; (2000 bright states), and last 2000 |mI=+1&gt; state (dark state) <br> These data are directly from Monte Carlo simulation and does not consider detection efficiency, which is added later by add_detection_efficiency following binomial distribution.

针对题为《机器学习增强的重复读出》的论文
本文件夹对应模拟的氮空位（NV）光动力学参数
training.mat文件包含data_plus（维度为5000×8000）、data_minus（维度为2500×8000）与data_zero（维度为2500×8000）三组数据，分别对应初始态为+1（暗态（dark state））、-1（亮态（bright state））与0（亮态（bright state））的14N核自旋体系（注：+1为暗态）
test.mat文件包含data_test数据（维度为4000×8000），其中前1000条样本对应|m_I=-1⟩态，中间1000条对应|m_I=0⟩态（共计2000个亮态样本），最后2000条对应|m_I=+1⟩态（暗态样本）
上述数据均直接来源于蒙特卡洛（Monte Carlo）模拟，未考虑探测效率；探测效率后续通过add_detection_efficiency函数基于二项分布（binomial distribution）添加