Anomaly detection using an ensemble of multi-point LSTMs

Abstract

As technologies for storing time-series data such as smartwatches and smart factories become common, we are accumulating a lot of time-series data. With the accumulation of time-series data, the importance of time-series abnormality detection technology that detects abnormal patterns such as Cyber-Intrusion Detection, Fraud Detection, Social Networks Anomaly Detection and Industrial Anomaly Detection is emerging. In the past, time-series anomaly detection algorithm has evolved mainly from univariate data. However, with the development of technology, time-series data has become complicated, and the corresponding deep learning-based time-series anomaly detection technology has been actively developed. Currently, most industries rely on deep learning algorithms to detect time-series anomalies.

In this paper, we propose an anomaly detection algorithm with an ensemble of multi-point LSTMs that can be used in 3 cases of time-series domains. We propose our anomaly detection model in 3 steps. The first step is a model selection step, in which a model is learned within a user-specified range, and among them, models that are most suitable are automatically selected. In the next step, collected output vector from M LSTMs is completed by stacking ensemble techniques the previously selected models. In the final step, anomalies are finally detected using the output vector of second step. We experiment with comparing the performance of the proposed model with the state-of-the-art time-series detection deep learning model using 3 real world datasets. Our method shows robust accuracy and F1 score in three datasets.