Click here to access the slides
Words, Documents and Distance: Deep Learning and Semantic Analysis
In the last few years algorithms developed by Google have given researchers powerful new tools to investigate textual data. Using so called “deep learning” neural network techniques it is possible to model the relationships between words, sentences and documents using numerical values to represent the content of the text. The paper presents an overview of the word2vec and doc2vec technologies Google have developed. It then shows examples of semantic mapping using open ended text data. An example of “semantic text clustering” is also shown, this is the application of traditional cluster analysis applied to text using numeric representations obtained from the doc2vec algorithm.