Real-Time Sentiment-Based Anomaly Detection in Twitter Data Streams

We propose an approach for real-time sentiment-based anomaly detection (RSAD) in Twitter data streams. Sentiment classification is used to split the data into independent streams (positive, neutral, and negative), which are then analyzed for anomalous spikes in the number of tweets. Four approaches for evaluating the data streams are studied, along with the parameters that adjust their sensitivity. Results from an evaluation show the effectiveness of a probabilistic exponentially weighted moving average (PEWMA) coupled with a sliding window that uses median absolute deviation (MAD).