A Novel LSSVM Based Algorithm to Increase Accuracy of Bacterial Growth Modeling

Area of Science:

• Microbiology
• Computational Biology
• Bioinformatics

Background:

• Advanced algorithms are revolutionizing predictive microbiology, particularly bacterial growth modeling.
• Accurate prediction of bacterial growth is crucial for various applications in food safety and biotechnology.

Purpose of the Study:

• To develop a more accurate hybrid algorithm for predicting bacterial growth curves.
• To enhance predictive microbiology studies through improved bacterial growth modeling.

Main Methods:

• Bacterial growth of *Listeria monocytogenes* and *Escherichia coli* was modeled using sigmoid functions (Logistic, Gompertz) and Least Square Support Vector Machine (LSSVM).
• Parameters 'c' and 'σ' for LSSVM were optimized using the non-dominated sorting genetic algorithm-II (NSGA-II), creating the NSGA-II-LSSVM hybrid algorithm.
• Model performance was evaluated using Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE).

Main Results:

• LSSVM demonstrated more precise bacterial growth modeling compared to traditional sigmoid functions.
• The NSGA-II-LSSVM hybrid algorithm exhibited superior predictive accuracy over the standard LSSVM method.
• Optimized parameter estimation using NSGA-II significantly enhanced prediction capabilities.

Conclusions:

• The NSGA-II-LSSVM hybrid algorithm provides superior bacterial growth prediction by accurately estimating model parameters.
• This approach enhances the predictive potential of LSSVM in microbiological studies.
• The study highlights the effectiveness of hybrid algorithms in advancing predictive microbiology.