失真测量的分析--参数估计和展开
Analysis of distorted measurements -- parameter estimation and unfolding
摘要：
1.从失真的测量参数推断进行了讨论。2.模糊的测量值在没有显式正则化的情况下展开。相应的结果是无偏的，并允许拟合参数和应用定量拟合优度检验。3.常见的展开方法（迭代EM与早期停止，截断SVD，ML适合曲率，熵和范数处罚）进行了测试和比较，彼此调整正则化参数，以尽量减少在所有情况下的综合平方误差（ISE）。除了直方图表示，样条近似被认为是。所有的模拟表明，EM方法导致更小的ISE比竞争的方法。特别有前途的是EM展开为样条逼近。这些研究是基于不同的分布，事件数量，分辨率和足够的独立模拟，以获得结论性的结果。建议用EM方法将数据展开为b样条近似，并用直方图和误差矩阵来补充结果，所述直方图和误差矩阵是通过展开为相对宽的箱而获得的，而不需要显式正则化。
英文摘要：
1. Parameter inference from distorted measurements is discussed. 2. Smeared measurements are unfolded without explicit regularization. The corresponding results are unbiased and permit to fit parameters and to apply quantitative goodness-of-fit tests. 3. Common unfolding methods (iterative EM with early stopping, truncated SVD, ML fits with curvature, entropy and norm penalties) are tested and compared to each other with the regularization parameter adjusted to minimize the integrated square error (ISE) in all cases. Apart from histogram representations, spline approximations are considered. All simulations indicate that the EM method leads to smaller ISEs than the competing approaches. Especially promising is the EM unfolding to spline approximations. The studies are based on different distributions, event numbers, resolutions and enough independent simulations to obtain conclusive results. It is proposed to unfold data with the EM method to b-spline approximations and to supplement the results with histograms and error matrices obtained by unfolding into relative wide bins without explicit regularization.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
