Semantic Search engines

Use your time for information, not for search

Unlike in the web the relevance of search results in the enterprise-context can't be measured by the grade of links of a document, because company intranets provide much lower link-structures.The more important becomes the semantic analysis of the documents that have to be indexed with help of knowledge models and text mining. Like this the system can understand easier, which contents are within the document. It also learns, which terms, phrases or entities (places, persons, products, branches, etc.) of a company are related.

The Semantic Web Company as your partner

Nowadays a powerful semantic enterprise search-engine considers the following factors:

  • Data integration: Connectivity with existing systems like databases, portal systems, CMS, DMS or enterprise wikis
  • Data amount: Scalability and high performance also with huge user access
  • Application integration: Flexibility regarding integration of search services in third-party applications and GUI-customising
  • Web connectivity: Powerful web-crawling for integrating information from the web into the intranet search
  • Relevancy ranking: Intelligent algorithms to measure the relevance of a document regarding the search query

What's the effort good for?

By using controlled vocabularies as a base of your enterprise search system we conduct you to the semantic age of information management. Following functions could simplify your daily life very soon:

Auto-complete: The system completes the search query automatically and suggests search terms that might be of interest for the user. With auto-complete spelling mistakes can be avoided and search time can be minimised. If the user forgot about the right spelling of a term, he will be supported by auto-complete to "remember" the right term.

Faceted search: Facets can be compared with dimensions of a search-room like places, persons, document types, time period, topic, etc. With each mentioning of a facet the list of results shrinks and the results show high precision.

Moderated search/ Refining search: The user looks for "Enterprise Search" and gets thousands of results. Terms which appear often in the results, are suggested to refine the search like "Sharepoint Enterprise Search" or "Information Retrieval". The results can be limited fast and a targeted search through many documents is possible.

Automatic search expansion: This function (also: query rewriting) enlarges the search query with related terms. An example: the user looks for "Enterprise Search", the search query is enlarged by synonyms like "corporate search", both terms are linked by OR, thereby mostly multilingual document repositories can be scanned more precise.

Similarity Search: For example one document was found and the system suggests similar documents from the repository or while one creates a document, existing matching text fragments are suggested. Like this, existing resources are used more effective. This function is used for example within knowledge management applications, while identifying similar "Lessons Learned", for example.

Person search/ Search for contact persons: Modern search technologies can extract automatically person names from huge document repositories, or you could search: "Where was Mr. Miller mentioned together with Mr. Jones?". Like this you can create expert search engines that help identifying contact persons for specific topics within your organisation precisely.

Tagging/ Tag Recommender: First, a document is analysed by the search engine for generating automated keywords (tags). Those are suggested to the employee, like this the tagging remains simple and high grade. The search index will be be strongly enriched because mainly fitting terms are indexed together with the document.