On the identification of verbs in computer programs of physiological models

Area of Science:

• Computer Science
• Computational Mathematics
• Software Engineering

Background:

• Mathematical models are often complex, making their underlying computational logic difficult to understand.
• Identifying the core actions (verbs) within source code is crucial for model interpretability.
• Existing methods may not adequately capture the structural essence of modeled actions.

Purpose of the Study:

• To present a rationale and computational techniques for identifying verbs (modeled actions) in mathematical model source code.
• To establish a method for programs to self-explain their computational processes.
• To link source code structures to relational database concepts for enhanced model understanding.

Main Methods:

• Utilizing flow graph methodology to represent program source code as a directed acyclic graph (DAG).
• Identifying complete loops within the DAG's disjoint intervals as structures equivalent to modeled verbs.
• Mapping these verb structures to functional dependencies in the intensional form of a relational database.

Main Results:

• Source code can be systematically decomposed into intervals representing modeled actions (verbs).
• These verb structures are uniquely mappable to relational database functional dependencies.
• Simulation is conceptualized as database retrieval, and logical inferences are derived via relational algebra.

Conclusions:

• Verb definitions derived from code structure are essential for self-explaining programs.
• This approach provides a formal framework for understanding mathematical models through computational logic.
• The proposed methodology bridges source code analysis with relational database theory for enhanced model interpretability.