7 ms·
1) API for classifying article into topics, or alternatively extracting the entities from a piece of text. Basically NLP as a service. 2) Twitter automation: S
by AznHisoka 13y ago
1) API for classifying article into topics, or alternatively extracting the entities from a piece of text. Basically NLP as a service.
2) Twitter automation: Sort of like IFTTT. Let me specify rules on when to automatically retweet an user.
- agibsonccc 13y agoI'm working on a comprehensive scraping tool for this. See: http://agibsonccc.github.io/blog/apidocs.html http://agibsonccc.github.io/blog/apidocs.html for the baseline functionality. The service will work as follows: Select some urls Pick what you want: (People,Organizations,Locations,Books,Intervals,Dates,Emails ,phone numbers, among lots of other ones) Hit go and download the data to a spreadsheet (each column is based on the types you specify). The system is fairly accurate not only due to Machine Learning, backed by a huge type mapped search index based on dbpedia/freebase (other data sources are added later) The core right now is to focus on making web based information retrieval simple (export to a spreadsheet) I'm open to other parts of the system as well. For example, I did experiments with the API to do sentiment analysis based on a comment thread or picking out reviews. Note: backend is done. I'm mainly working on refinements and a bit of tuning of the NER engine (web pages suck compared to normal articles) Get at me if you're interested, I'd love to get some feedback. Email is in my profile.