Electronic Health Records (EHRs) are an important source of information that till now is only partially explored. This happens because a substantial part of EHRs is in the form of free text that needs to be processed by text mining techniques to extract relevant information, and such techniques are still under development. A similar remark can be done for Information Extraction from biomedical literature. In this case, the large amount of scientific literature produced every year makes it difficult to remain updated even in restricted domains.
Recognition and classification of the pharmacological substances, chemical-disease relation extraction, chemical-chemical relation from the biomedical data are of great interest to the physicians and researchers. Automatic extraction of such information from the biomedical data, that contains mostly unstructured data, are substantial.
Application of the Natural language processing (NLP) in the biomedical data to extract the information from the unstructured data is very successful. NLP has been widely adopted in many fields due to its better achievement towards the unstructured data. NLP can extract the concept even if it is described in some other way. It captures the information from the text and presents it to the user.
However, NLP still facing limitations such as contextual ambiguities, synonyms, ill-formed sentences in the biomedical data. These limitations need to be addressed so that the extracted information from biomedical data can further be processed to achieve the desired result.


Curriculum: Computer Science/Engineering and Control Systems

Tutor: Alfonso E. GEREVINI, Ivan SERINA
Relatore: Alberto Lavelli (FBK)

email: t.mehmood@unibs.it


Link to the research group web page

Link to publication list