Discrimination Power of Polynomial-Based Descriptors for Graphs by Using Functional Matrices

Area of Science:

• Graph theory
• Mathematical chemistry
• Cheminformatics

Background:

• Graph-theoretical matrices are crucial for encoding molecular structures.
• Previous work utilized polynomial-based descriptors and the Randić matrix for graph encoding.
• Generalizing and improving upon existing graph-theoretical matrix methods is an ongoing challenge.

Purpose of the Study:

• To investigate the discrimination power of graph measures derived from graph-theoretical matrices.
• To generalize and extend previous research on polynomial-based descriptors and the Randić matrix.
• To introduce and validate a new functional matrix approach for enhanced graph discrimination.

Main Methods:

• Exploration of graph-theoretical matrices for structural information encoding.
• Generalization of existing methods, specifically referencing work using the Randić matrix.
• Application of a novel functional matrix approach to a set of exhaustively generated graphs.

Main Results:

• Demonstration of enhanced discrimination power compared to previous methods.
• Unique identification of exhaustively generated graphs using the new functional matrix approach.
• Validation of the proposed method's superiority over polynomial-based descriptors.

Conclusions:

• The functional matrix approach offers superior graph discrimination capabilities.
• This method provides a more unique encoding of structural information in graphs.
• The findings advance the field of graph-theoretical analysis for structural representation.