ontology-Semantic BaSed Search With oc | miner速
Who iS ontochem? We are a text and knowledge mining IT solution provider, based in Halle Germany and founded in 2005. Our clients are large and small life science companies and publishing houses that want to extract data and knowledge from proprietary and open access document collections. Our crossdisciplinary team of informatics engineers, biologists, pharmacologists and chemists is dedicated to provide high quality and cost-effective solutions that no other company can provide. comPany miSSion Our mission is to become a leader in document based data mining and analysis in the life science industry – utilizing proprietary text mining and analysis algorithms as well as domain specific ontologies and databases. With our technologies we are annotating and analyzing all available patents, scientific and related domain literature and websites to support our client’s scientific work and decision making.
What we do Named entity recognition: Ontology based text parsing annotates known entities such as proteins, genes, chemicals, diseases, organisms, anatomy terms of organisms and plants, physiological terms, person and company names, geographical and geopolitical locations, but also quantitative information about time and other measures. Ontologies: text mining results are only as good as the underlying ontologies. Therefore we build our own ontologies, that are larger, more accurate, and support natural language processing. Relationship detection: Subsequent semantic text analysis identifies a large variety of more complex explicit relationships between terms such as for example physiological effects of chemical compounds. Even more advanced are our tools to extract implicit relationships – thus enabling hypothesis building and new discoveries. Searching: The extracted terms and relationships are stored in proprietary condensed No-SQL databases and can be searched in the intranet or internet by conventional web-browser search interfaces.
Ontology-Semantic Based Searching Disambiguation of homonyms: mole = animal, skin abnormality, chemical unit Natural language processing: quercetin reduces the risk of certain cancers = anticancer compound synonym & relational searches SjÜgren’s syndrome = autoimmune disease
What we offer annotation and analysis of your documents with our ontologies building domain specific ontologies for your problem area custom intranet and internet search engine and browser solutions solving your scientific problem with our IT resources, e.g. review on state-of-the-art, proposing new solutions and hypothesis based on literature evidences
How we do it OC|miner速 is our modular and scalable text processing engine that is using a patented technology for very large and ultrafast dictionaries, containing more than 200 million different terms from different knowledge domains. Shallow parsing and natural language processing together with semantic rule collections is used to extract complex relationships. Custom build and graphical user web-browser back-ends, including chemical search engines allow a tailor made presentation and retrieval of data.
PROBLEM
OC|MINER速
SOLUTION USER INTERFACE OUTPUT
Level 1: Text Mining find “named entities” Example: find documents with “apple” (and malus domestica Borkh., alma, etc.)
INPUT
OC|MINER®: read, analyze, organise
Level 2: Data Mining find facts or relationships between entities Example: find the concentration of quercitin in apple
Level 3: Knowledge Mining find complex goal oriented relationships Example: what is the most suitable plant to extract quercitin as a food additive for the treatment of diabetes?
Why choose us With OC|miner® text mining is both faster and more accurate than using competing solutions. Our quality assurance allows for precision rates of more than 95% and recall rates higher than 90%. Processing very large document collections only takes hours and searching complex relationships is done within parts of seconds. Ontology based searching allows to search for concepts – e.g. the search term “plants” will retrieve all documents that mention any plant name in more than 10 languages. Relationship searching allows hypothesis searching, either explicit or implicit – e.g. is there a natural compound in a plant extract that could be used as food additive that is not patented so far.
get in touch with us Please write us at info@ontochem.com or call us at +49 345 4780470 for more information. We would be pleased to arrange a WebEx conference or send you more supporting information on our technologies and capabilities. Ontochem GmbH Heinrich-Damerow-Strasse 4 06120 Halle Germany info@ontochem.com www.ontochem.com