Load Characteristics Screening of Electric Vehicle Charging Stations Based on LASSO Regression

In recent years, the number of electric vehicles has continued to increase, and in the future, electric vehicles will be widely connected to the power grid. Electric vehicles are not only a flexible electricity load, but also a distributed energy storage resource. Their charging and discharging behavior will have an undeniable impact on the power grid. Studying their load characteristics is crucial for the operation of the power grid. In this paper, the maximum mutual information coefficient MIC was used to quantify the load characteristics and correlation of EV charging stations. On this basis, the influencing factors of EV charging load were considered and the influencing factors of EV charging load were screened by LASSO regression model. The example analysis shows that: the current temperature, the load value of the previous hour, and the load value of the same time one week before the current moment have a prominent impact on the load prediction of EV charging stations. The results indicate that Lasso based feature selection for electric vehicles is more reliable.