Skip to main content
Fig. 3 | Algorithms for Molecular Biology

Fig. 3

From: A novel method for inference of acyclic chemical compounds with bounded branch-height based on artificial neural networks and integer programming

Fig. 3

ac An illustration of Phase 1: a Stage 1 for preparing a data set \(D_{\pi }\) for a graph class \({\mathcal {G}}\) and a specified chemical property \(\pi \); b Stage 2 for introducing a feature function f with descriptors; c Stage 3 for constructing a prediction function \(\psi _{{\mathcal {N}}}\) with an ANN \({{\mathcal {N}}}\); de An illustration of Phase 2: (d) Stage 4 for formulating an MILP \({{\mathcal {M}}}(x,y,g;{\mathcal {C}}_1,{\mathcal {C}}_2)\) and finding a feasible solution \((x^*,g^*)\) of the MILP for a target value \(y^*\) so that \(\psi _{{\mathcal {N}}}(x^*)=y^*\) (possibly detecting that no target graph \(G^*\) exists); (e) Stage 5 for enumerating graphs \(G^*\in {\mathcal {G}}\) such that \(f(G^*)=x^*\)

Back to article page