Apache Spark Deep Learning Cookbook

How it works...

Consider the following:

• There are various methods to measure the semantic similarity between words. The one we are using in this section is based on cosine similarity. We can also explore linear relationships between words by using the following lines of code:

 def nearest_similarity_cosmul(start1, end1, end2):    similarities = got2vec.most_similar_cosmul(        positive=[end2, start1],        negative=[end1])start2 = similarities[0][0]print("{start1} is related to {end1}, as {start2} is related to         {end2}".format(**locals()))return start2

• To find the cosine similarity of nearest words to a given set of words, use the following commands:

nearest_similarity_cosmul("Stark", "Winterfell", "Riverrun")nearest_similarity_cosmul("Jaime", "sword", ...