The hosted project is a naive bayes document classifier. Currently using a multinominal model, however the Bernoulli and Ensemble model are also planned. The project was initially written using VBA in excel (I know the horror!), to assist in my PhD research. Whilst it functioned and was okay, even if I do say so myself, the single threaded nature of Excel was a bottle neck (link to VBA project for anyone interested). Plus all the bells and whistles of an OOP langauge were sorely missed, So I rewrote it in VB.net. I can hear the C# boys and girls scoffing and spitting out their coffee. You're wasting good java there people. Why do we always focus on the things that divide us? We are all brothers and sisters in code 😎 .
- Indicating model accuracy
- Saving project and dependencies, for easier quick tests.
- Removing connectives
- User Guide
- Direct integeration with twitter api, for sentiment analysis of tweets
- PDF to text conversion
- Export model as JSON
- Optimise Model
- Application Overview
- Models
- Multinominal
- Bernoulli
- Ensemble
- Project Walk Through
- Project Setup
- Selecting drive for project folder structure
- Setting folder path and folder alias
- Adding preclassified documents (Training Set)
- Adding validation documents (Test Set)
- Train Model
- Project Setup
- Model Accuracy
- Navigating UI and Settings
- Examples and Use Cases
My contact details are below, if you wish to discuss anything, reach me at:
- Twitter: http://www.twitter.com/kevin_fasusi
- Google+: http://google.com/+KevinFasusi