As the reform of the electricity market progresses, emerging power market entities, including load aggregators and virtual power plants, facilitate market participation for small-scale users, enhancing benefits and demand-side response. Efficient mining and load forecasting of user electricity consumption are crucial for these entities to formulate aggregation strategies and make pricing decisions. To this end, from the perspective of load aggregation management of massive small-scale users, this research firstly adopts the interval value method to characterize the user load sequence, then introduces the interval k-means and hierarchical clustering methods to classify and identify the user clusters with similar electricity consumption patterns. The interval prediction model is established to realize probabilistic prediction of massive user loads, with a view to providing effective decision support and reducing transaction risks for load aggregators, virtual power plants and other subjects to participate in the market. Finally, the case analysis utilizing actual electricity consumption data verifies that the interval clustering analysis improves the accuracy of load forecasting, and the ability of the forecasting model to fit the nonlinear interval-valued load sequences affects the effectiveness of the forecasting algorithm framework.