Figure1

Figure 1. Domain knowledge-guided interpretive ML strategy. A: Feature selection with $$ \mathfrak{R}_{\phi_{j}} $$ and data screening through $$ 3\sigma $$ criterion. B: Novel oxidation Cr equivalent of FM steels derived from joint contributions of elements. C: A prototype "divide-and-conquer" algorithm, TCLR, proposed for capturing the influences of features on time exponents and activation energy. D: Formulas of crucial terms derived by the SISSO under the constraint of prior domain knowledge.