Month: February 2017

Automatically Extract Contextually Accurate Keyphrases

Using advanced linguistic focused Artificial Intelligence and Machine Learning processes found in to the patented Extractor technology, the xAIgent RESTful web service provides subscribers with an effortless software service for automatically extracting contextually accurate keyphrases / key words / key terms – from any subject matter content.
The xAIgent RESTful service, uses the patented Extractor hybrid Artificial Intelligence and machine learning Linguistic Technology to provide subscribers with the most accurate and contextually relevant key terms from any subject domain text, automatically (unsupervised).
In contrast, it’s worthwhile to note there are other keyphrase extraction systems and most based on heuristic and Bayesian derived key word extraction models. Each inherently requiring their systems to be manually trained per each subject domain the developer / user wishes to employ.  Training is a process whereby a library (corpus) of pre-defined, domain specific content and keywords must first be compiled an then incorporated in to the comparative structure (supervised process) of that system.  Cumbersome at best. Tedious and time intensive expert knowledge required.
The xAIgent automatic keyphrase extraction RESTful web service is ready for consumption immediately, without further training, supporting English, French, German, Japanese, Korean and Spanish, and provides subscribers with the most accurate, contextually relevant keyphrases of any solution available today.
Where would an automatic, contextually accurate key word / keyphrase extraction RESTful service be useful?
Think of document management and content management systems, where their contents must first have key terms / key words assigned to each document prior to its inclusion in the repository. If the author has not previously tagged the content, then a subject matter expert must be employed to appropriately determine the key terms that describe the document. Read / Re-read the content, identify the key words, terms and phrases and then annotate to the document. Then the document / content can be included into the document management system.
To help alleviate the read / re-read document process, often document management systems will have a generic list of subject terms to select from and assign to the document. That may be all well and good if the documents being consumed are all of similar subject matter, but is that really the best approach? Wouldn’t it be better to have contextually accurate key terms per document that would then allow the true value of the documents being included into the management system to be exposed? Allowing them to b effectively accessed, searched, referenced and reported on.
Of course and the simple answer… subscribe to the xAIgent RESTful web service and have objective (human generated key words by their nature are subjective), contextually accurate key terms generated automatically. For all unstructured content. In other words, set your own xAIgent enhanced system to work through a collection of content folders and have each document automatically associated with its own set of key words / key terms / key phrases / tags – Automatically. Come back when the process is completed (over night) and start to fully realize the enhanced value that has now been surfaced for that collection of documents.
There are many other aspects of the xAIgent (Extractor) service to note and we’ll do that in subsequent editions, including why xAIgent is so good at retrieving key phrase content from websites and why research shows the automatic xAIgent (nee Extractor) key term extraction process carries an accuracy rating of up to 87 percent.