Enhancing Cookie Formulations with Combined Dehydrated Peach: A Machine Learning Approach for Technological Quality...

Area of Science:

• Food Science
• Machine Learning Applications
• Nutritional Analysis

Background:

• Cookie quality is influenced by ingredients and processing methods.
• Dehydrated fruits can enhance nutritional and sensory properties of baked goods.
• Predictive modeling aids in optimizing food product development.

Purpose of the Study:

• To predict and optimize cookie quality parameters using Artificial Neural Network (ANN) and Support Vector Machine (SVM) models.
• To investigate the impact of dehydrated peach and pretreatment methods on cookie attributes.
• To determine optimal formulation for enhanced cookie quality.

Main Methods:

• Development of separate ANN and SVM models for 32 distinct cookie quality parameters.
• Application of sensitivity analysis to identify key influencing factors.
• Multi-objective optimization using ANN models to find ideal formulation parameters.

Main Results:

• ANN models achieved high prediction accuracy (r² up to 1.000) for diverse quality parameters.
• SVM models showed strong performance (r² up to 0.981) for specific variables.
• Dehydrated peach and osmotic pretreatment significantly influenced cookie composition and quality.

Conclusions:

• ANN and SVM are effective tools for predicting and optimizing cookie quality with fruit enrichment.
• Osmotic pretreatment enhances the benefits of lyophilized peach in cookie formulations.
• An optimal formulation of 15% osmotically pretreated lyophilized peach was identified for superior cookie quality.