A multi-source transfer learning method and device for air conditioner energy consumption prediction in the face of scarce samples and electronic equipment

By using a multi-source transfer learning framework and model-level regularization functions, the problem of sample scarcity in air conditioning energy consumption prediction is solved, achieving high-precision prediction under different building and climate conditions, and improving the model's transfer accuracy and generalization ability.

CN118966450BActive Publication Date: 2026-06-09SOUTHEAST UNIV +3
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN ยท China
Patent Type
Patents(China)
Current Assignee / Owner
SOUTHEAST UNIV
Filing Date
2024-08-06
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In air conditioning energy consumption prediction, existing technologies suffer from insufficient sample data, especially for newly built or significantly renovated buildings. Insufficient initial data or data gaps lead to insufficient samples, resulting in overfitting of training network parameters. A single transfer learning strategy is unable to cope with the distribution differences under different building types and climatic conditions, thus affecting the prediction results.

Method used

A multi-source transfer learning framework is adopted. By introducing a model-level regularization function into the objective function, the differences between multiple source domains are quantified, the parameters of the transfer learning model are dynamically adjusted, the base prediction model is trained using multiple source domain datasets, and the transfer model is built in the target domain. The weights of the regularization term are adjusted by combining gradient descent optimization and cross-validation to achieve knowledge sharing and model generalization.

Benefits of technology

It improves the accuracy and generalization performance of air conditioning energy consumption prediction under scarce sample conditions, provides an effective solution, and provides a reference for large-scale coordinated and normalized adjustment of building flexible loads.

โœฆ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118966450B_ABST
    Figure CN118966450B_ABST
Patent Text Reader

Abstract

The present application relates to a kind of in the case where available sample data is scarce, using multi-source transfer learning method is carried out air conditioner energy consumption prediction method, belong to the application field of data mining technology in energy management system.Its method is: data is preprocessed;According to the data for pre-processing multi-source transfer learning model determines loss function;According to the distribution difference between source domain and target domain dynamic adjustment transfer learning model's parameter;Using transfer learning model is carried out air conditioner energy consumption to target domain, obtains energy consumption prediction value.The present application proposes a kind of multi-source transfer learning method for facing the rare sample air conditioner energy consumption prediction, by increasing model level regular function in objective function, realize the knowledge sharing between multi-source domain and improve the transfer precision and generalization performance of model, provide effective solution for the air conditioner system energy consumption prediction under the sample scarce scene.
Need to check novelty before this filing date? Find Prior Art