A robust ability to perform Natural Language Processing is at the heart of every document capture and data extraction project. Natural Language Processing (NLP) refers to using computers to process human language. As opposed to a giant library that needs to be fully present to perform NLP activities, Grooper uses NLP tools across its platform. Grooper extractors, regular expressions, optical character recognition (OCR), and implementation of the TF/IDF algorithm all involve the processing of natural language to achieve best-in-class results for even the most complicated uses cases.
Grooper supplements core NLP capabilities with the ability to natively process text as n-grams and via porter stemming, in addition to supporting configurations that implement more complex NLP methods, such as sentiment analysis, part-of-speech tagging, named entity tagging, and feature tagging for classification, among others.
To clarify, the main difference between a standard NLP library, like the Stanford Library, and the Grooper implementation of NLP, is that Grooper uses NLP throughout the product, not just as an add-on to a single piece of the product. NLP, and other advanced ML/AI functionality, is embedded throughout Grooper.
Grooper NLP capabilities offer the following advantages: A robust ability to process text that comes from degraded documents using our state-of-the-art suite of image processing tools. If you can’t OCR the text, your NLP project cannot begin. No- or low-code configuration, depending on complexity: Grooper offers users a robust design studio that allows for most problems to be solved without the use of code; for particularly difficult applications, Grooper allows for code to be inserted in single-statement “expressions,” or through building out functionality using our extensive SDK.
Intelligible AI: All Grooper core AI and NLP algorithms are fully transparent, with tools available to designers to understand how extraction and classification decisions are being made. AI decisioning can be inspected at every point and adjusted how the user sees fit.
A “human-first” approach to problem solving: Grooper core philosophy is that human intelligence and human methods of understanding and parsing documents are better than opaque machine approximations; as such, Grooper is a tool geared at encoding and automating human understandings and decisions. Grooper architects are taught to ask the question: “How would I know what to do here” when approaching difficult problems; our software is built around this approach, and Grooper solutions are built to work with architects and end-users, rather than against them.
We look forward to having the opportunity to explain and demonstrate these features in more detail.