SCIMPULSE.DE
Updated 28 days ago
MER (Maximum Entropy of Runs 1 ) is a novel statistical parameter-estimation method which gains the sample-information from the measurement-clusters, below and beyond positioned along the model-curve. The distances between the measurement-points and model-curve no more play a role. Formally these clusters are iterations2 ("runs") of unilaterally to the model-curve lying measurements. As clusters they are accumulations (unbroken sequences) of same-quality-objects, here the measurement values. Objective function (criterion) of the method is the probability of that up- and down-run system. This probability reflects the entropy of the ensemble . The by parameter-variation gained maximum probability, is the MER-solution and appears as a p-dimensional space wherein different values of the identical dimension are equivalent. The solution of our 'symmetration' process mostly has exactly one value .- With the MER- technique the common residual-based techniques (OLS, ML, etc.) become -..