Scholarlink

This chapter presents a study to predict the generation and consumption of electric energy in Mexico. This kind of study’s importance lies in identifying possible areas of opportunity to improve distribution by state organizations. For this purpose, machine learning algorithms based on time series were used to represent the behavior of the energy market in Mexico for generation and consumption based on historical public data spanning from 1980 to 2022. The models considered for the analysis were SARIMA, Holt-Winters, and VAR; in all cases, a training and test set was used in a proportion of 80-20. To conclude the viability of each model used, the AIC and BIC criteria and the predictions made for each of the models under analysis were used. The results showed that the VAR model offers better conclusions for predicting generation and consumption trends for the Mexican electricity sector for the next 10 years.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting Electricity Generation and Consumption Through a Time Series Algorithm in Mexico

  • 章节

  • 发表日期：

    2026年

查看全文
  • Miguel Arturo Gallegos Acuña,
  • Luis Carlos Méndez-González,
  • Luis Alberto Rodríguez-Picón,
  • Isidro Jesús González-Hernández,
  • Iván Juan Carlos Pérez Olguín,
  • Luis Asunción Pérez Dominguez

摘要

This chapter presents a study to predict the generation and consumption of electric energy in Mexico. This kind of study’s importance lies in identifying possible areas of opportunity to improve distribution by state organizations. For this purpose, machine learning algorithms based on time series were used to represent the behavior of the energy market in Mexico for generation and consumption based on historical public data spanning from 1980 to 2022. The models considered for the analysis were SARIMA, Holt-Winters, and VAR; in all cases, a training and test set was used in a proportion of 80-20. To conclude the viability of each model used, the AIC and BIC criteria and the predictions made for each of the models under analysis were used. The results showed that the VAR model offers better conclusions for predicting generation and consumption trends for the Mexican electricity sector for the next 10 years.