PoolParty Components Involved in Recommendation - An Overview

The recommendation functionality is an emergent feature of the PoolParty Semantic Suite platform. At the heart are the PoolParty services that process data from various sources, use graph databases and indexing engines, offer tools for end users, and provide the foundation for customer applications. PoolParty Semantic Recommender allows users to make sophisticated matches between defined concept with a focus not just on similarities between items but rather more on context and background. The PoolParty Recommender can retrieve results not obviously related to the input but rather through ontology-based classifications and relations within a knowledge graph.

Let us now take a look at the interaction between the PoolParty components in a recommendation process.

The PoolParty Thesaurus is responsible for the management of the knowledge model and provision of such to other components including the PoolParty Extractor and PoolParty Recommender. The PoolParty Extractor is used by the Thesaurus to enrich the knowledge model based on information from texts and documents, and by PoolParty UnifiedViews in data processing pipelines for content annotation and also in the PoolParty Recommender to annotate content from user's input.

The PoolParty UnifiedViews on the other hand serves the creation of a knowledge model as well as Recommender database by transforming, linking and integrating data sources of various types, whereas the PoolParty GraphSearch is the data backend of the Recommender. Recommender's frontend is the Recommender Workbench where the subject matter expert can configure and set up the recommender. Moreover, the PoolParty Recommender API serves as an interface for integrators who want to make use of the PoolParty Recommender.