Correlation analysis of factors affecting wind power based on machine learning and Shapley value

Abstract

An analysis of the impact of various factors on wind power can help grid dispatchers understand the characteristics of wind power output and improve the accuracy of wind power forecasting. A correlation analysis method of factors affecting wind power is proposed based on machine learning and the Shapley value. First, factors affecting wind power and the method of constructing wind power models based on machine learning are introduced. Then, to measure the influence of factors on wind power, the Shapley value is proposed based on the wind power model. In addition, calculation methods, properties, and application scenarios of the Shapley value are introduced. Finally, based on the actual data of a wind farm, the method is used to analyse environmental factors affecting wind power, and the main factors affecting the wind farm are determined. The experimental results show that the method can identify important factors affecting wind power and measure the complex non-linear relation between each environmental factor and wind power.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

A data‐driven approach to state assessment of the converter valve based on oversampling and Shapley additive explanations

“…The basic idea of SHAP is to utilize an additive model g ( x ) to fit the trained classifier f ( x ), as shown in formula (5) [26]: $$\begin{equation}f({\bm{x}}) = g({\bm{x}}) = {\phi _0} + \sum_{i = 1}^n {{\phi _i}} \end{equation}$$where n is the number of characteristics, ϕ 0 is the prediction reference value of the model, that is, the mean value of all samples state levels, and ϕ i is the SHAP value of the i ‐th characteristic.…”

“…This paper combines SHAP to provide a feasible interpretation scheme for the output of complex CatBoost models. The basic idea of SHAP is to utilize an additive model g(x) to fit the trained classifier f(x), as shown in formula (5) [26]:…”

Multivariate Data-Driven Models for Wind Turbine Power Curves including Sub-Component Temperatures

“…An anticipation of the results which are of interest in the present context is that the covariates should be ranked not only by how much the average error diminishes when each covariate is included (as is done in this work), but also for how much the average error (once the set of covariates is selected) depends on each variable. The latter information can be obtained by computing the Shapley coefficients [38,39] for each variable and, for the data sets of this study, the rotational speed ranks as the highest, which means it is the most explanatory. In Table 2, the determination coefficients between the rotor speed ω and the temperatures selected for the SVR regression are reported.…”

Polynomial surface-fitting evaluation of new energy maximum power generation capacity based on random forest association analysis and support vector regression

“…(1) The feature selection model utilizes the RF algorithm to analyze the correlation of key variables that influence wind power generation capacity, such as meteorological equipment, and selects the two most significant correlation variables. Chuanjun et al (2021) did not consider the wind power time-varying characteristics and complex nonlinear relationships as well as the utilization of a gradient boosting decision tree and artificial neural network supervised learning algorithms in the original correlation analysis to train the wind power model. A correlation analysis based on the importance of the influencing factors and the amount of bias dependence was proposed.…”