YZYuzhuo Zhang

Yuzhuo Zhang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Yuzhuo Zhang has a total of 10 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Ultra-Short-Term Load Dynamic Forecasting Method Considering Abnormal Data Reconstruction Based on Model Incremental Training

    by Guangyu Chen, Yijie Wu, Li Yang, Ke Xu, Gang Lin, Yangfei Zhang, Yuzhuo Zhang

    In order to reduce the influence of abnormal data on load forecasting effects and further improve the training efficiency of forecasting models when adding new samples to historical data set, an ultra-short-term load dynamic forecasting method considering abnormal data reconstruction based on model incremental training is proposed in this paper. Firstly, aiming at the abnormal data in ultra-short-term load forecasting, a load abnormal data processing method based on isolation forests and conditional adversarial generative network (IF-CGAN) is proposed. The isolation forest algorithm is used to accurately eliminate the abnormal data points, and a conditional generative adversarial network (CGAN) is constructed to interpolate the abnormal points. The load-influencing factors are taken as the condition constraints of the CGAN, and the weighted loss function is introduced to improve the reconstruction accuracy of abnormal data. Secondly, aiming at the problem of low model training efficiency caused by the new samples in the historical data set, a model incremental training method based on a bidirectional long short-term memory network (Bi-LSTM) is proposed. The historical data are used to train the Bi-LSTM, and the transfer learning is introduced to process the incremental data set to realize the adaptive and rapid adjustment of the model weight and improve the model training efficiency. Finally, the real power grid load data of a region in eastern China are used for simulation analysis. The calculation results show that the proposed method can reconstruct the abnormal data more accurately and improve the accuracy and efficiency of ultra-short-term load forecasting.


  • Open Access

    Intelligent Identification and Order-Sensitive Correction Method of Outliers from Multi-Data Source Based on Historical Data Mining

    by Guangyu Chen, Zhengyang Zhu, Li Yang, Wenhao Huang, Yuzhuo Zhang, Gang Lin, Shengjie Zhang

    In recent years, outliers caused by manual operation errors and equipment acquisition failures often occur, bringing challenges to big data analysis. In view of the difficulties in identifying and correcting outliers of multi-source data, an intelligent identification and order-sensitive correction method of outliers from multi-data sources based on historical data mining was proposed. First, an intelligent identification method of outliers of single-source data is proposed based on neural tangent kernel K-means (NTKKM) clustering. The original data is mapped to high-dimensional feature space using Neural Tangent Kernel, where the features of outliers are acquired by K-means clustering to realize the accurate identification of outliers. Second, an order-sensitive missing value imputation framework for multi-source data (OMSMVI) was proposed. The similarity graph of sources with missing data was constructed based on multidimensional similarity analysis, and the filling order decision was transformed into an optimization problem to realize the optimal filling order decision of missing values in multi-source data. Finally, a neighborhood-based imputation (NI) algorithm is proposed. Based on the traditional KNN filling algorithm, neighboring nodes of sources with missing data are flexibly selected to the achieve accurate correction of outliers. The case experiment was operated on actual power grid data, and the results show that the proposed clustering method can identify outliers more accurately, and the determined optimal imputation sequence has higher accuracy, which provide a feasible new idea for the identification and correction of outliers in the process of data preprocessing.