Research on Guangxi Smart Substation Project Investment Prediction and Influencing Factors Based on XGBoost-SHAP

As an important component of the power system, substations undertake the crucial task of ensuring the safe operation of the power grid, and the annual construction investment is huge. Controlling the investment in substation engineering construction can effectively reduce the costs of power grid enterprises and maximize the efficiency of resource allocation. Based on the feasibility study investment data of Guangxi Power Grid’s smart substation projects in the past three years, this paper conducts an exploration. The principal component analysis method is used to select 11 main features that affect the investment level of substations, and the k-means clustering algorithm is utilized to optimize the sample data set. An index system of influencing factors for the investment in Guangxi smart substation projects is constructed, and an investment prediction model for smart substation projects based on XGBoost is established. Through case studies, the effectiveness of XGBoost in substation investment prediction is verified. Finally, combined with the SHAP values, the influence degree of each feature on the prediction model is obtained, which verifies the validity and applicability of the prediction model. It provides an important reference basis for reasonably controlling the investment in Guangxi substation projects and also offers strong support for investment decision-makers in making investment decisions regarding substation engineering construction.