Novelty Detector For Categorical Data
thatDot, Inc.
Product Overview
Ingest any kind of categorical data through a REST API for a real-time score of its novelty. Immediately find anomalies that are hiding in a high-volume stream of data. Each observation is evaluated for novelty, using context to reducing false positives, and providing an explanation of what makes the observation so unusual.Real-time ingestion of streaming or batch data-even combined batch and streaming.Categorical data: strings, names, numbers, file paths, etc. No need to convert to numbers.Get more than just the anomaly score, get an explanantion why it's anomalous!Visualize your dataset to understand common trends and outliers.Great for small data of dozens of observatons, all the way up to large data streams in the trillions.Run many different anomaly detection streams through the same container, each separated into their own context.Deployed as a container for easy integration into applications.Choose how sensitive the anomaly detection should be using well-distributed scores.Simple operation and management.Complete and interactive API documentation is included in the container.Turn high-volume data into high-value data in real-time with thatDot Streaming Anomaly Detector.
Version: 0.12.0
By
thatDot, Inc.
Categories: ML Solutions
Business Intelligence & Advanced Analytics
Operating System
Linux
Delivery Methods
Container
Supported servicesAmazon ECSAmazon EKS
Highlights: An API for Real-time Novelty Detection on Non-Numeric Data Contextual Anomaly Detection to Dramatically Reduce False Positives Each API call returns rich details about your observation: novelty score, probability, Information Content, and more. Beyond just a score, you can understand WHY the observation is so unusual.
Category: Infrastructure Software
Delivery Method: Container
Company Information
Company Name: thatDot, Inc.
About Company: thatDot unlocks the value of streaming data, making it accessible and actionable in real-time. Find patterns and anomalies in events, logs and metrics, to trigger actions.
Anomaly Detector is the world's only real-time categorical anomaly detector. File names, IDs, labels, etc. are fed in and anomalies are separated from routine and highly cardinal data, reducing Big Data to insight.
Connect triggers real-time actions from high-volume streaming data. Events, logs, and metrics stream in, are automatically connected, and trigger analysis, ETL, anomaly detection and automation.