Title | Anomaly Detection of Power Big Data Based on Improved Support Vector Machine |
Publication Type | Conference Paper |
Year of Publication | 2022 |
Authors | Yu, Jinhe, Liu, Wei, Li, Yue, Zhang, Bo, Yao, Wenjian |
Conference Name | 2022 4th International Academic Exchange Conference on Science and Technology Innovation (IAECST) |
Date Published | dec |
Keywords | anomaly detection method, Big Data, composability, Data models, equivalent vector, False Data Detection, feature extraction, Human Behavior, Power big data, pubcrawl, Real-time Systems, Reliability engineering, resilience, Resiliency, Support vector machines, Technological innovation, Vector machine |
Abstract | To reduce the false negative rate in power data anomaly detection, enhance the overall detection accuracy and reliability, and create a more stable data detection environment, this paper designs a power big data anomaly detection method based on improved support vector machine technology. The abnormal features are extracted in advance, combined with the changes of power data, the multi-target anomaly detection nodes are laid, and on this basis, the improved support vector machine anomaly detection model is constructed. The anomaly detection is realized by combining the normalization processing of the equivalent vector. The final test results show that compared with the traditional clustering algorithm big data anomaly detection test group and the traditional multi-domain feature extraction big data anomaly detection test group, the final false negative rate of the improved support vector machine big data exception detection test group designed in this paper is only 2.04, which shows that the effect of the anomaly detection method is better. It is more accurate and reliable for testing in a complex power environment and has practical application value. |
DOI | 10.1109/IAECST57965.2022.10062053 |
Citation Key | yu_anomaly_2022 |