Sensor-Based Evaluation of Purslane-Enriched Biscuits Using Multivariate Feature Selection and Spectral Analysis.

Area of Science:

• Food Science and Technology
• Agricultural Science
• Sensory Science

Background:

• Functional ingredients offer health benefits beyond basic nutrition.
• Purslane (Portulaca oleracea L.) stalk flour is a potential source of valuable nutrients.
• Optimizing ingredient incorporation requires robust quality assessment methods.

Purpose of the Study:

• To develop a sensor-integrated framework for evaluating purslane stalk flour in butter biscuits.
• To determine the optimal substitution level of purslane stalk flour for enhanced product quality.
• To demonstrate a non-destructive method for quality monitoring in food manufacturing.

Main Methods:

• Design of Experiments (DoEs) with multisensor probes (EC, pH, TDS, ORP) and digital imaging.
• Multivariate analysis including Repeated Relief Feature Selection (RReliefF) and Principal Component Analysis (PCA).
• Regression modeling and linear programming for optimization.

Main Results:

• Reduced 54 measurements to 19 informative features using RReliefF and PCA.
• Identified an optimal purslane stalk flour substitution of 9.62% using regression and linear programming.
• Biscuits with 9.62% flour showed improved texture, enhanced mineral content (Ca, Mg, Fe, Zn), stable color, and maintained sensory appeal.

Conclusions:

• The sensor-based framework provides reliable, low-cost, non-destructive quality assessment.
• Data-driven optimization enables enhanced nutritional profiles and desirable product characteristics.
• This approach supports efficient product development and quality control in the food industry.