Abstract: Embodiments are directed to extracting specified data items from narrative text. In one scenario, a computer system accesses narrative textual information which includes data items that are to be identified and extracted. The computer system identifies specified data items in the narrative textual information that are to be extracted from the narrative textual information. The computer system then filters the identified data items to remove false positive identifications. The false positive filtering includes classifying the identified data items as specified data items, so that classified data items are identified as true positive items that are to be extracted from the narrative textual information. The computer system further extracts, from the narrative textual information, those filtered data items that were classified as being true positive items.

Abstract: Embodiments are directed to extracting specified data items from narrative text. In one scenario, a computer system accesses narrative textual information which includes data items that are to be identified and extracted. The computer system identifies specified data items in the narrative textual information that are to be extracted from the narrative textual information. The computer system then filters the identified data items to remove false positive identifications. The false positive filtering includes classifying the identified data items as specified data items, so that classified data items are identified as true positive items that are to be extracted from the narrative textual information. The computer system further extracts, from the narrative textual information, those filtered data items that were classified as being true positive items.