Segmenting users into various groups not only provides you with insights about your community, but also enables you to create personalized content for them, e.g. in form of advertisements, tutorials, support etc.
In contrast to an Unsupervised Learning approach, with Supervised Learning you’ll actually know what your segments are made of. To do so, however, you need large amounts of labeled data to train a classifier. This is where kern comes into play, where you can integrate information sources of various types to scale your data labeling. For instance, you can embed labeling functions that declare typical characteristics of advocate users - this information is integrated with various other sources using Weak Supervision. This ensures high quality data, which you can easily further manage to build great classifiers and gain insights into your development.
If you’re interested, give our free version a try. We’re there to help and guide you along the way if you have any questions.