Identification of Related Factors of Users' Power Consumption and Prediction Model of Power Consumption Based on Random Forest Algorithm

Given the characteristics of big data associated with power consumption prediction, such as numerous types, vast volume, high dimensionality, and rapid generation speed, we propose a method for subspace clustering analysis of extensive user electricity characteristics after studying the evaluation i...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2023 International Conference on Internet of Things, Robotics and Distributed Computing (ICIRDC) s. 205 - 210
Hlavní autoři: Zhu, Zheng, Xiao, Shuang, Zhang, Ruijia, An, Bailong
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 29.12.2023
Témata:
Big Data
Correlation
Electricity
Identification of related factors
power consumption forecast
Power demand
Prediction algorithms
Predictive models
Random forest algorithm
Simulation
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Given the characteristics of big data associated with power consumption prediction, such as numerous types, vast volume, high dimensionality, and rapid generation speed, we propose a method for subspace clustering analysis of extensive user electricity characteristics after studying the evaluation indicators of user electricity traits. This method aims to extract various power consumption patterns among users. Users are categorized into groups based on their different power consumption patterns, and a mutual information matrix is utilized to identify factors related to power consumption within these user groups. These factors encompass regional and industry economic data, climatic conditions, and electricity prices, leading to the development of a big data prediction model for power consumption. This approach effectively discerns pertinent factors influencing power consumption across various user groups and mitigates the adverse effects of disparities in power consumption models on prediction accuracy. Simulation results validate the high predictive accuracy of this method, making it suitable for extensive big data analysis and processing.
DOI:10.1109/ICIRDC62824.2023.00043