mCLESS: The Multi-Class Least-Error Square Sum for Interpretable Classification
Hwan Hee Park, Seung Heon Lee, Hwamog Kim, Seongjai Kim · 2023
Some machine learning algorithms are considered as black boxes, because the models are sufficiently complex and they are not straightforwardly interpretable to humans. Lack of interpretability in predictive models can undermine trust in those models in many application areas. The article introduces a new interpretable machine learning algorithm, called the Multi-Class Least-Error Square Sum (mCLESS). It is linear, simple to implement, and interpretable. Its nonlinear expansion is discussed. This simple algorithm turns out to be superior to many popular machine learning algorithms. Various experimen-tal results involving synthetic datasets and UCI datasets are given to verify the claim.