Anomaly Detection of High-Frequency Sensing Data in Transportation Infrastructure Monitoring System Based on Fine-Tuned Model

Anomaly detection has been widely studied in previous studies during recent decades; however, there are still some challenges for high-frequency sensing data. The most challenging task is to deal with a large volume of data in an extremely short time. In previous studies, it has been proved that converting the data into pictures can improve the speed of anomaly detection. However, the training of the image recognition algorithms, such as deep learning models, still needs a long time. Fortunately, the fine-tuned convolutional neural networks (CNNs) give us opportunities to detect anomalies in high-frequency data quickly. Thus, a four-stage model is proposed for anomaly detection in high-frequency data. Using a real-world dataset, one designed CNN, four widely used fine-tuned CNN, and two popular machine learning methods are compared by the confusion matrix. Moreover, three ensemble methods are proposed to improve the accuracy of the detection. The results show that the majority of voters can improve the overall accuracy by 2.09%. Significantly, the accuracy of minor and outlier classes can be increased by ensemble learning which is a challenge in practice.

Automated summary

Anomaly detection in high-frequency sensing data is challenging due to the large volume of data and limited time. This paper proposes a four-stage model for quick detection using fine-tuned CNNs and compares it with other algorithms. The results demonstrate the effectiveness of ensemble methods in improving overall accuracy as well as accuracy for minor and outlier classes.