Construction of a machine learning-based prediction model for rice varieties-cooking-mastication

Area of Science:

• Food Science and Technology
• Machine Learning Applications in Food Processing
• Sensory Science

Background:

• Understanding the link between rice characteristics, cooking, and oral processing is crucial for food product development.
• Existing methods for predicting oral processing outcomes are limited.
• A multi-scale approach is needed to integrate various factors influencing rice consumption.

Purpose of the Study:

• To establish a machine learning-driven predictive framework for quantifying relationships between rice physicochemical properties, cooking conditions, and oral processing.
• To develop and compare various machine learning models for predicting mastication and bolus properties.
• To enable end-to-end prediction from cooking inputs to oral outcomes for optimized rice product design.

Main Methods:

• Characterized 56 rice varieties for texture, morphology, and amylose content.
• Measured in vivo mastication parameters (chewing time, number of chews, saliva volume) and bolus properties (particle size, reducing sugar content).
• Developed and evaluated five prediction models: multiple linear regression, extreme gradient boosting (XGBoost), support vector machine (SVM), long short-term memory network (LSTM), and convolutional neural network.

Main Results:

• Hardness, gumminess, and chewiness were key determinants of mastication behavior (r > 0.7).
• Grain length-breadth ratio negatively influenced bolus particle size (r > 0.7).
• LSTM achieved optimal mastication prediction (test R² = 0.9499), SVM excelled in bolus property forecasting (R² = 0.9168), and XGBoost captured cooking-texture interactions (R² = 0.8161).

Conclusions:

• A machine learning framework can effectively predict rice mastication and bolus properties based on physicochemical characteristics and cooking conditions.
• Cascade models (XGBoost-LSTM, XGBoost-SVM) enable comprehensive prediction from cooking to oral processing.
• This framework advances digital optimization of rice processing and sensory quality, facilitating tailored food design.