Select Page

Search and Recommendation

Get relevant results in less time with sophisticated search and recommendations.

Trigger a full-text search query and narrow down the results using search facets.

Get further reading suggestions and dive deeper into your topic.

Easily configure your search and recommendation settings in the PoolParty Workbench.

Graphwise Search and Recommendation Components comprise a range of search capabilities built on Graph AI. These applications are highly configurable, allowing the user to display information according to their specific needs. With Graphwise Search and Recommendation, users can benefit from knowledge retrieval that understands the user’s natural language, where faceted search helps to narrow down results, and recommendations are based on intent and context.

Get sophisticated content recommendations from a tailor-made recommender system.

Graphwise recommender systems are driven by domain, context, and intent, which means that they are highly customizable to an organization’s specific use case. With a Graphwise Recommender, content that is not explicitly related (but still relevant) to the search query can be surfaced; a user can find exactly what they are looking for PLUS benefit from additional helpful information.

Avoid the “cold-start problem” by using structured data and text annotations

Use semantic footprinting and query expansion to get recommendations that are not only based on similarity but implicit connections

Benefit from traceable recommendations to make sound decisions

Incorporate linked data such as Wikidata and/or DBpedia to easily broaden the knowledge model

Factor in multilingualism with bundled concepts

Easily set up your configuration in the Workbench.

The Graphwise Workbench makes it easy to configure your search or recommendation space in the backend. A system administrator or user with editing permissions can turn features on/off with a simple click and decide which factors they want the applications to draw from. These configurations can easily be pushed to a Graphwise ADF (Application Development Framework) frontend.

Choose the Taxonomy and the graph repositories you want to run the search space on – you can draw from multiple projects at once

Determine which facets you want to display in the search frontend and how deep in the hierarchy you want the facets to go

Enter the minimum score a concept has to reach to be extracted. The score ranges from 1 to 100 where a higher score means that the concept is more relevant for the processed text 

Resolve language ambiguities by drawing upon the connected thesaurus and local context – what would normally take sophisticated coding work, only takes seconds by toggling a lever 

Use a corpus and shadow concepts to broaden the scope of the recommendations

Set the languages for multilingual search and recommendations

While it is typically a data scientist or machine learning expert who handles the configuration of such systems, the user-friendly interface of Graphwise Workbench can even be used by a subject matter expert, who is typically a non-technical user.

Take a deep dive with an expert.

Please fill out the form to get connected with a Graphwise Team Member. Get ready for your guided walkthrough and full-access account!