TextTeaser, an application that uses Natural Language Processing and Machine Learning to automatically summarize text, has been open sourced.
The best way to appreciate the effectiveness of TextTeaser is try it. You can either copy and paste some text into the window it provides or just enter the url of an article - feel free to try it with one of our news items.
Once it has speedily created a point-by-point summary you can share a link to it, either as text or an image, or you can use embed code to add it to a webpage.
(click to enlarge)
As its co-creator Jolo Balbin explained in an interview for Programmable Web, he started work on the algorithm as part of his graduate studies and he is very keen to convert it into a real world application. As part of his research he did tests to compared his algorithm to others that performed automatic summarization and found that it performed better. As it uses machine learning in combination with NLP (Natural Language Programming) its performance will improve as more articles are summarized.
TextTeaser's API and its documentation is now on Mashape, with both Freemium and subscription options. For those who want to get involved with the project the code is on GitHub under an MIT license. It requires the sbt build tool and use of the Scala IDE is recommended.