(a) selecting n input variables I¹, I², . . . Iⁿ (n is an integer), each input variable corresponding to a structural property or an electronic property of one or more ground state model structures GS^A1, . . . GS^Ap (p is an integer) and a plurality of transition state model structures TS^A1, TS^A2, . . . TS^Am (m is an integer) associated with the one or more ground state model structures,

wherein each of the one or more ground state model structures GS^A1, . . . GS^Ap and each of the plurality of transition state model structures TS^A1, TS^A2, . . . TS^Am are derived from one or more first training heteroatomic ligand-metal compound complexes, each complex comprising a first training heteroatomic ligand;

(b) assigning a quantitative value to each n input variable I¹, I², . . . Iⁿ, for each of the ground state model structures GS^A1, . . . GS^Ap and each of the transition state model structures TS^A1, TS^A2, . . . TS^Am,

(c) determining, by at least one processor of a device, the relative energies of each of the ground state model structures GS^A1, . . . GS^Ap and each of the transition state model structures TS^A1, TS^A2, . . . TS^Am;

(d) generating a machine learning model based upon correlating the quantitative value of each n input variable I¹, I², . . . Iⁿ with the relative energies of each of the ground state model structures GS^A1, . . . GS^Ap and each of the transition state model structures TS^A1, TS^A2, . . . TS^Am,

(e) identifying, based on the machine learning model, one or more of the n input variables I¹, I², . . . Iⁿ associated with [1] the difference in energies between one of the ground state model structures GS^A1, . . . GS^Ap and at least one of the plurality of transition state model structures TS^A1, TS^A2, . . . TS^Am [ΔG(TS−GS) or ΔΔG(TS-GS)] or [2] the difference in energies between any two or more of the plurality of transition state model structures TS^A1, TS^A2, . . . TS^Am [ΔG(TS-TS) or ΔΔG(TS-TS)];

(f) generating, based upon the one or more n input variables I¹, I², . . . Iⁿ identified from step (e), a first target heteroatomic ligand-metal compound complex for olefin oligomerization and comprising a first target heteroatomic ligand, wherein the first target heteroatomic ligand-metal compound complex is characterized by n output variables O¹, O², . . . Oⁿ, each having a quantitative value corresponding to a structural property or an electronic property of one or more ground state model structures GS^B1, . . . GS^Bx (x is an integer) or any of a plurality of transition state model structures TS^B1, TS^B2, . . . TS^By (y is an integer) associated with the one or more ground state model structures,

wherein each of the one or more ground state model structures GS^B1, . . . GS^Bx and each of the plurality of transition state model structures TS^B1, TS^B2, . . . TS^By are derived from the first target heteroatomic ligand-metal compound complex, each complex comprising a first target heteroatomic ligand;

(g) identifying one or more performance parameters associated with an olefin oligomerization reaction and the value of the performance parameters for the one or more first training heteroatomic ligand-metal compound complexes and the first target heteroatomic ligand-metal compound complex; and

(h) repeating steps (a)-(f) one or more times using the quantitative values of the n output variables O¹, O², . . . Oⁿ of the first target heteroatomic ligand-metal compound complex as an input dataset of new n input variables I^1.1, I^2.1, . . . I^n.1 derived from one or more second training heteroatomic ligand-metal compound complexes for olefin oligomerization and comprising a second training heteroatomic ligand, which is computationally evaluated against the machine learning model to generate a second target heteroatomic ligand-metal compound complex comprising a second target heteroatomic ligand, wherein the second target heteroatomic ligand-metal compound complex is characterized by quantitative values of an output dataset of new n output variables O^1.1, O^2.1, . . . O^n.1, and having one or more second target heteroatomic ligand-metal compound complex performance parameter values.