Simulating soft data to make soft data applicable to simulation

Area of Science:

• Systems Biology
• Computational Biology
• Biomedical Informatics

Background:

• Biomedical processes are influenced by subjective, 'soft' data, which is currently underutilized in simulations.
• Existing simulation models often exclude non-crisp or subjective measures, limiting their comprehensive applicability.
• This research addresses the gap by developing a method to integrate soft data into simulations.

Purpose of the Study:

• To introduce an input generator for soft data (input generator SD) to make subjective measures applicable in biomedical simulations.
• To demonstrate the utility of soft data in enhancing the accuracy and scope of systems biological models.
• To explore the application of machine learning and regression techniques for simulating non-crisp processes.

Main Methods:

• Development of an input generator for soft data (input generator SD).
• Application of machine learning approaches to simulate odour intensity ratings.
• Utilization of standard regression techniques for data analysis and simulation.

Main Results:

• The input generator SD successfully makes soft data applicable for simulation purposes.
• Simulations using machine learning and regression techniques yielded satisfactory performance.
• The generated results are comparable to those produced by the simulated system itself.

Conclusions:

• Soft data should be integrated into systems biological simulations to improve model realism.
• The input generator SD provides a viable method for incorporating subjective measures into simulations.
• Machine learning and mathematical approaches can effectively model non-crisp processes for modifying biological models.