GEOSS post 1: Semantic spaces of textual metadata content

Sometimes we can discover more information in metadata abstracts than in all other fields, especially when we have so many records as GEOSS can provide. This global data sharing architecture boasting having 300 million metadata records on datasets and services is pretty much operational and delivering data on daily basis. Yet, nobody knows really what is the global picture. In the following series of posts we will try to discover the contents with you. 

Below you can find a picture showing neural network that learned word embedding of words in the texts describing the data. Out of curiosity we have picked up some themes and mined the neural network for finding the path between these terms. The results are telling us not only how similar the terms are but most importantly, how their contextual semantic spaces interact and eventually, what the GEOSS is all about at all.

Leave a Comment