Sense Analysis Service
The Sense Analysis service annotates text with senses. It is based on a combination of several machine learning algorithms and rules-based components. It handles well-capitalized documents, poorly-capitalized documents (e.g., tweets, informal documents) and Internet search queries. The generated senses can be correlated to corresponding entities in other well-known namespaces (e.g., Wikipedia).
Concepts
The following documents describe the key concepts for using the service.
- Sense Annotation Concepts: Describes the key concepts for understanding the output of sense analysis.
- Using the Sense Analysis Service: Describes how to use the service.
- Understanding the Semdoc Format: The easiest format for retrieving the senses is the "semdoc" (semantic document) XML format. This document explains the format using an example.
- Understanding the Confidence Thresholds: Applications must use thresholds to avoid using unreliable results. This document explains them.
- Sense Analysis Recipes: Document processing can take several seconds (even minutes for very long documents). Many applications can advantageously trade a little result precision for increased speed and reduced cost against your profile.
- Document Processing Scheduling: Describes how jobs are scheduled and what an application can do to obtain better performance.
Getting Started
- Create an account on this web site or log in: Login
- Create a project: Create Project
- Follow the example of your choice: Ruby
- Check out the API: text/disambiguate
