Anomaly Detection in Time Series data using Hierarchical Temporal Memory Model

In this data-driven world, the Anomaly detection technique plays a key role in various domains. Since the amount of generated data is huge, conventional anomaly detection techniques using batch processing are inefficient since the cost of storing and processing large amounts of data incurs various costs. Therefore, the best algorithm for this task should be able to detect anomalies in a real-time manner and with no human intervention. In this paper, we discuss the application of the Hierarchical Temporal Memory algorithm to detect anomalies in real time and in an unsupervised manner. We applied this algorithm to the stock market dataset and analyzed the performance. We also applied the algorithm on an artificially created dataset with a known point anomaly. This work tests the performance of the algorithm for real-world applications and detects anomalies.