Time Series Anomaly Detection; Detection of anomalous drops with limited features and sparse examples in noisy highly periodic data

TL;DR

This paper explores machine learning and statistical methods to detect anomalous drops in noisy, periodic traffic data streams with limited labeled examples, focusing on sustained anomalies rather than short-term fluctuations.

Contribution

It introduces a combined approach using TensorFlow-based regression models and rule-based detection to identify sustained anomalies in noisy, periodic time series data.

Findings

Intersection of detection methods improves accuracy

Models perform well on periodic data streams

Some data streams are non-periodic, limiting model effectiveness

Abstract

Google uses continuous streams of data from industry partners in order to deliver accurate results to users. Unexpected drops in traffic can be an indication of an underlying issue and may be an early warning that remedial action may be necessary. Detecting such drops is non-trivial because streams are variable and noisy, with roughly regular spikes (in many different shapes) in traffic data. We investigated the question of whether or not we can predict anomalies in these data streams. Our goal is to utilize Machine Learning and statistical approaches to classify anomalous drops in periodic, but noisy, traffic patterns. Since we do not have a large body of labeled examples to directly apply supervised learning for anomaly classification, we approached the problem in two parts. First we used TensorFlow to train our various models including DNNs, RNNs, and LSTMs to perform regression and predict the expected value in the time series. Secondly we created anomaly detection rules that compared the actual values to predicted values. Since the problem requires finding sustained anomalies, rather than just short delays or momentary inactivity in the data, our two detection methods focused on continuous sections of activity rather than just single points. We tried multiple combinations of our models and rules and found that using the intersection of our two anomaly detection methods proved to be an effective method of detecting anomalies on almost all of our models. In the process we also found that not all data fell within our experimental assumptions, as one data stream had no periodicity, and therefore no time based model could predict it.