 |
Models in Psydex represent People, Companies, Places, Trades or anything that can be defined with words, symbols, attributes and value dimensions.
Semantic Topic Models apply to Unstructured data and Data Topic Models define well Structured data such as market trading, transactions or packet data. Models are uniquely identified by a Topic Symbol and customizable attributes such as classifications. Semantic Models are Universal - Source and Index Independent.
Semantic Topic Models are represented by expressions or semantic rules comprised of characters, words and phrases and operators that define relative or temporal proximity to other words and phrases. Expressions are defined using the Psydex Query Language (PQL). Topic Models are managed and stored completely separate from Indexes. This approach allows models to evolve as language, taxonomies and ontologies evolve independent of
Indexes
Semantic Topic Models are independent of Indexes and can be updated and viewed in a historical context in real time
Data Topics are based on a declarative model for defining multiple dimensions for a given information source.
The dimensions represent the classes of things a Data Topic model can be declared with and utilize.
Data Topic models are also defined by the Aggregate Functions (e.g. Count, Sum, Average True Range (ATR) ) applied to a specified dimension, and the criteria used to select the appropriate records.
For example, a Data Topic may be the volume-weighted average price (VWAP) for the December 2010 West Texas Intermediate (WTI) crude oil futures contract.
All data topics are more generally just topics, and as such, are processed by the services offered by the Psydex Analytics Grid, such as Statistics Service.
|
|
|
...
|
|
