Rozanski Rafal, Kawecka Elzbieta, Perec Andrzej
Faculty of Technology, Jacob of Paradies University, 66-400 Gorzów Wielkopolski, Poland.
Materials (Basel). 2025 Jul 2;18(13):3131. doi: 10.3390/ma18133131.
The paper proposes a new coefficient assessing the classification ability of parameters. In contrast to previously used indices, it does not require data normalization, examines the correlation between parameters with the highest classification ability, and determines, based on this, a complementary set that enables effective differentiation of surfaces that differ significantly. The empirical part is based on the values of 83 parameters that characterize the stereometric features of 22 surfaces created through different machining processes.
本文提出了一种评估参数分类能力的新系数。与先前使用的指标不同,它不需要数据归一化,能检验具有最高分类能力的参数之间的相关性,并据此确定一个互补集,该互补集能够有效区分差异显著的表面。实证部分基于83个参数的值,这些参数表征了通过不同加工工艺创建的22个表面的立体特征。