thatDot Novelty performs anomaly detection without requiring predefined patterns or data labeling. It creates a detailed behavioral fingerprint for strong contextual awareness as data streams in, scoring every event based on its unusualness in the dataset.
02
Categorical Data Analysis
The system can analyze categorical data as-is, even with high cardinality, without the need to transform it into sparse numeric data. This capability is particularly useful for entity and user behavior analytics (EUBA), network address and traffic analytics, and digital twin simulations.
03
Rapid Processing
thatDot Novelty can process over 15,000 events per second on an 8-CPU, 16 GB virtual machine, outperforming the Isolation Forest algorithm which processes 500 events per second under similar conditions.
04
Built-In Data Transformation
The system includes built-in data transformation functions that can be applied directly in the stream, eliminating the need for a separate transformation layer.
05
Contextual Anomaly Scoring
thatDot Novelty scores events based on novelty, uniqueness, information content, and probability within seconds of analyzing existing data.
Your plan caps how many capabilities are shown — upgrade to see the full list
