How the classification service works?
There are two ways to interface with our service:
- Use the web interface from any development language (interface is described here: web interface)
- Use a DLL provided by us that manages requests and cache.
Integration with Komodia’s SDK
How do you know what is the category of the site?
There are number of universal ways to classify a site, most of them involved some sort of Machine learning and SVM algorithms based on the keywords in the site, in addition we leverage public resources and manual human review.
Which languages do you support?
We fully support the English version, and have partial support for: German, French, Spanish, Italian, Dutch, Japanese, Chinese, Arabic, Croatian, Czech, Finnish, Greek, Hebrew, Norwegian, Polish, Romanian, Portuguese, Russian, Swedish, Turkish.
Partial supports means that only the main categories are supported (about 20) with strong focus on porn/adult sites, for example a site that would be “science” in Chinese will be classified as “General – Not porn” since science is not a main category.
Do you have real time virus detection?
At the moment we don’t offer security services in regards to viruses, we do have plans to integrate real time phishing database.