TriniData: our technologies
Knowledge is profit!

Semantic technologies

All our work is focused on the use of semantic technologies. It is a set of specifications defined as W3C standards, defining data modeling principles and representation syntax (OWL), graph database query language (SPARQL), etc. Despite its "web" origin, these technologies can be effectively used in corporate automation. Their use offers the next technological opportunities:

1. For the data integration:
1.1 Usage of the information model for storing unified corporate data structure, not depending on information representation in particular systems.
1.2 Usage of this model for organizing centralized data exchange within single IT ecosystem, or between many systems.
1.3 Use of graph database for storing data model and master data allows to break the relational model restrictions, namely:
- Ability to assign several types (classes) to the single object, representing its various aspects - for example, Customer and Organization.
- Dynamic attribute model inheritance from all its classes and superclasses.
- Ability to assign several values to the single object property (if allowed by constraints).
- Ability to add new classes and attribute definitions, not affecting existing data.
1.4 Ability to implement model-driven software, which automatically adjusts its behavior according to the model changes, not requiring any customization.
1.5 Ability to implement model-driven information exchange architectures, in which all the components are easily replaceable, that prolongs system's lifecycle as a whole.

2. For the data analysis purposes:
2.1 Ability to represend and examine qualitative properties of the objects, their structural, functional relationships in the form of logical axioms.
2.2 Ability to analyse causal, logical relationships between events.
2.3 Ability to use automatic inference functionality for producing new knowledge and automating analysis.
2.4 Implementation of the search engines using logical inference to obtain completely reliable, logically proven results for any queries formulated in the terms of the conceptual model.
2.5 The use of controlled natural language for data formalization and extraction.
2.6 Binding of heterogenous, disparate data, and extracting new value from it.

3. For the optimization solutions building purposes:
3.1 The complex, heterogenous, dynamically updateable arrays of source information are required for the fully functional optimization solutions development. These arrays may be effectively gathered and updated using integration solutions based on semantic technologies.
3.2 Semantic technologies simplify creation of simulation models, particularly - the multi-agent environments, due to possibility to define the rules of objects interaction in the form of the set of exioms. These axioms may be aumomatically executed during simulation process. Our simulation modeling platform is imlementing this principle.
3.3 Information model makes it easy to support and update models during their operation through the single point.
3.4 Technologies are allowing to set the optimization goals correctly, without simplification, and to decompose them to the required details level.
3.5 The use of ontological basis for model conceptualization simplifies the work with it for a subject matter area expert, reducing the need in IT specialists.

From the cost/efficiency point of view, the use semantic technologies for analytic, optimization, integrationtask has the next advantages:
1. Decreasing the cost of development and support of IT systems.
2. Significant cost reduction for integration.
3. Increasing speed of IT systems adaptation to the requirements change.
4. Assuring new level of quality of analytical calculations, extending sphere of application for analytics.
5. Assuring knowledge accessibility, minimizing cost for its search, rise of efficiency of its use.
6. Increasing reliability, efficiency, safety of the industrial and other systems by using optimization algorithms of unrivaled quality.

Semantic data representation principles

Semantic technologies are using the next concepts to represent all information:

  • Class - groups objects by type. For example, there may be a class named "Customers". Class definition itself does not bear any rules by which objects may be assigned to this class.
  • Attribute - used to define property that members of some classes may have. For example, objects of classes "Humans" and "Animals" may have attribute "Age".
  • Relationships - is a form of attributes that are describing relationships between objects. For example, the objects of the class "Persons" may have a values for attribute "Works in Company". This attribute may have multiple values.
  • Instances, or individuals, are representing information on particular objects, and their attributes values. For example, the object named "John Smith" may be an instance of the class "Person".