Semantic Search is an advanced technology for optimizing search accuracy, which goes beyond the traditional keyword/phrase search. Instead, it aims to understand the relationships between the words, thus making a better sense of the searcher’s intent and the query context.
To achieve this, the structured and unstructured data is transformed into a more intuitive and responsive knowledge paradigm – the knowledge graph, which enables highly contextual and richly personalized results.
The Goal
One of the biggest Pharma companies in the world needed to build a semantic search tool that would enable its users to easily find relevant information across huge volumes of siloed structured and unstructured data-sources.
The solution in place could not handle this operation efficiently as finding historical data in different documents with the available tools and systems took significant time. There was also a high rate of repetitive errors, which came from the lack of proper knowledge sharing and use of historical data.
The Challenge
The Pharma company needed an intelligent industry-specific solution that provides:
- automatic categorization and semantic sectioning of complex documents;
- normalization of both structured and unstructured data towards ontology terms used by text analysis pipelines;
- fusing of structured and unstructured data into a knowledge graph;
- powerful semantic search user interface to enable seamless data exploration through the knowledge graph.
The Solution: A Smarter Semantic Search Tool for Better Knowledge Insights
The semantic search solution provided by Ontotext enables users to get better knowledge insights by interlinking various siloed content based on semantic rules.
It was challenged with 5 diverse use cases, which required deep analysis of the content structure, information extraction from unstructured content (Health Regulatory documents, SOPs, technical manuals, etc.) and building a targeted knowledge graph with ingestion of structured datasets.
Why Choose Ontotext?
Currently, Ontotext’s solution provides:
- intuitive semantic search based on auto-suggest of concepts from the knowledge graph;
- multidimensional semantic filtering;
- term proximity search and knowledge exploration capabilities;
- provenance of all extracted facts from the original source document (with highlights and navigation within the content);
- semantic vectors based similarity search (later included in Ontotext’s leading semantic graph database GraphDB) enabling automatic matching between documents or parts of documents.
All five use cases were successfully implemented and two of them were nominated for the next phase of adoption plan for semantic technology within the company.
Do you think this case resembles your particular needs?
