Method for classification and prediction of rapeseed flavour during storage based on machine learning

By constructing a rapeseed flavor classification and prediction (RFCF) machine learning model, and combining feature engineering and parallel design of multiple algorithms, the problem of difficulty in quickly classifying and predicting rapeseed flavor changes in traditional methods is solved. This enables efficient and accurate judgment of the rapeseed aging process and supports quality control during rapeseed storage.

CN122333409APending Publication Date: 2026-07-03HUAZHONG AGRI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HUAZHONG AGRI UNIV
Filing Date
2026-03-31
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Traditional methods are insufficient for quickly and accurately classifying and predicting flavor changes in rapeseed during storage, and cannot meet the needs of large-scale testing.

Method used

A machine learning model for rapeseed flavor classification and prediction (RFCF) was constructed, employing a three-level architecture of feature engineering, multi-algorithm parallelism, and meta-model integration. Combined with electronic nose technology, the model uses algorithms such as random forest, support vector machine, and LightGBM to process rapeseed flavor data and achieve aging time and variety classification.

Benefits of technology

It improves the prediction accuracy of aging status of different rapeseed varieties, overcomes the limitations of single models and the problem of sensor data redundancy, and realizes efficient and accurate judgment of rapeseed flavor changes, providing reliable support for quality control during rapeseed storage.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122333409A_ABST
    Figure CN122333409A_ABST
Patent Text Reader

Abstract

This invention relates to the field of rapeseed flavor quality, and more particularly to a machine learning-based method for classifying and predicting rapeseed flavor during storage. A machine learning model for rapeseed flavor classification and prediction is constructed. The input layer receives flavor data from at least three different storage times. A feature engineering enhancement layer captures complex nonlinear relationships. The core algorithm layer employs a multi-algorithm parallel architecture, including four basic algorithms: Random Forest, LightGBM, Support Vector Machine (SVM), and Bayesian Regression. The output layer generates prediction results, which can output variety classification predictions, aging time classification predictions, or regression predictions of aging degree according to application requirements. This invention identifies key flavor indicators through feature engineering and automatic feature selection techniques, eliminates redundant information in electronic nose data, and enhances effective feature expression. By integrating multiple learning strategies, it improves the prediction accuracy of aging states for different rapeseed varieties, overcoming the limitations of a single model and enhancing generalization ability.
Need to check novelty before this filing date? Find Prior Art