Reader small image

You're reading from  Graph Data Science with Neo4j

Product typeBook
Published inJan 2023
Reading LevelIntermediate
PublisherPackt
ISBN-139781804612743
Edition1st Edition
Languages
Tools
Right arrow
Author (1)
Estelle Scifo
Estelle Scifo
author image
Estelle Scifo

Estelle Scifo possesses over 7 years experience as a data scientist, after receiving her PhD from the Laboratoire de lAcclrateur Linaire, Orsay (affiliated to CERN in Geneva). As a Neo4j certified professional, she uses graph databases on a daily basis and takes full advantage of its features to build efficient machine learning models out of this data. In addition, she is also a data science mentor to guide newcomers into the field. Her domain expertise and deep insight into the perspective of the beginners needs make her an excellent teacher.
Read more about Estelle Scifo

Right arrow

Using a transductive graph embedding algorithm

As we stated in the preceding section, a transductive algorithm is characterized by the fact that it works only on a full dataset, meaning it won’t be able to make any predictions on new observations. But, as with the centrality or community detection algorithms we have already crossed in the preceding chapters, these algorithms can be useful in circumstances where your graph is not evolving too fast. The GDS library currently contains two such algorithms: Node2Vec and Fast Random Projection (FastRP). We’ll describe the principles and usage of the Node2Vec algorithm. The usage of the FastRP algorithm will be very similar.

Understanding the Node2Vec algorithm

The Node2Vec algorithm is derived from the DeepWalk algorithm. In order to understand DeepWalk, we also need to know about the Word2Vec and SkipGram models.

As you can imagine, Word2Vec is an embedding algorithm for words within texts. As for a graph, a text...

lock icon
The rest of the page is locked
Previous PageNext Page
You have been reading a chapter from
Graph Data Science with Neo4j
Published in: Jan 2023Publisher: PacktISBN-13: 9781804612743

Author (1)

author image
Estelle Scifo

Estelle Scifo possesses over 7 years experience as a data scientist, after receiving her PhD from the Laboratoire de lAcclrateur Linaire, Orsay (affiliated to CERN in Geneva). As a Neo4j certified professional, she uses graph databases on a daily basis and takes full advantage of its features to build efficient machine learning models out of this data. In addition, she is also a data science mentor to guide newcomers into the field. Her domain expertise and deep insight into the perspective of the beginners needs make her an excellent teacher.
Read more about Estelle Scifo