Creating a Word2vec neural net using Deep Learning for Java (DL4j)

Word2vec can be seen as a two-layer neural net that works with natural text. With its typical usage, the input for the algorithm can be a text corpus, and its output is a set of feature vectors for words in that corpus. Note that Word2vec is not, strictly speaking, a deep neural network as it translates text into a numerical form that deep neural nets can read and understand. In this recipe, we will see how we can use the popular deep learning Java library named deep learning for Java (from this point on, DL4j) to apply Word2vec to raw text.

How to do it...

1. Create a class named Word2VecRawTextExample:

           public class Word2VecRawTextExample { 
   

2. Create a logger for this class. The logger facility has already been included in your project, as you have used Maven to build your project:

           private static Logger log = 
             LoggerFactory.getLogger(Word2VecRawTextExample.class); 
   

3. Start creating your main...