Custom entity recognition is a cloud-based API service that applies machine learning intelligence to empower the building of models for custom-named entity recognition tasks. Also known as “entity chunking” or “entity extracting,” as well as named entity identification, this service is a sub-task of information extraction that locates and classifies named entities mentioned in unstructured text.
Such text is generated and culled in various forms, from email messages to survey responses and Powerpoint presentations to Word docs. Examples of unstructured text can include images, video, or audio files.
With a named entity ID, data is categorized into pre-defined subsets such as people’s names, locations, medical codes, monetary values, and quantities. Xetlink’s custom-named entity recognition feature can identify and categorize within the unstructured text—whether that’s people, places, organizations, or amounts, for example.
You can analyze documents and extract business-specific entities or product codes that fit your need, whether that's analyzing the Big Island's geography or the coastal erosion rate along the Louisiana bayou. IBM defines an "entity" as "a single unique object in the real world that is being mastered. Examples of an entity are a single person, product, or organization."
Countless law offices, financial firms, and businesses – from Bank of America to Dollar Tree – extract thousands of complex unstructured text sources daily, from legal agreements to bank statements. For example, manually removing mortgage application data (i.e., asking Joe in Accounting for the mortgage application) would take days. With Xetlink's custom-named entity recognition feature, the process is automated, slashing time and effort.