We can send you emails of updates if you'd like (one email a week at most). Just enter your email address below, click on 'GO' and we'll be in touch.
This demonstrates how you can take a bunch of texts and categorise them just according to their content. What you get out are documents clustered according to meaning rather than by keyword. We have other demonstrations: planned categorising where you define the categories and provide a description; and unplanned categorising, where we do all the graft and you just tell us how many categories you need.
|What is this page doing?||This spontaneously categorises content into however many categories you want (within reason: you cannot categorise 4 documents into 5 or more categories and categorising them into 4 categories is fairly obvious). You can define the number of categories you want at the top of the screen.|
|Any example data?||Just click on "Fill in some data for me!" at the top and it will populate the form with some example data for you, particularly 8 short descriptions of things: 4 of them are computer related and 4 are animal related. Note that one from each mentions Python (which is both a computer programming language and a snake). Ideally, Roistr should sort al 8 documents into 2 categories, one with the computer related documents, the other with all the animals.|
|What do the results mean?||The results here will just show you all these documents organised into the number of groups you requested.|
|How does it work?||Roistr's semantic relevance engine analyses each document's "meaning" in terms of a language map and produces a vector that represents the meaning of the document in terms of this map. With this vector, we can mathematically compare each document against others. Once we've drawn a vector for each document, we just perform a basic k-means cluster analysis which provides the groupings.|