Graphs defined by topological information are helpful in many machine-learning scenarios. They can be used for community detection, node influence, classification, and other tasks. The performance a machine-learning mode can achieve on these tasks will strongly depend on the graph’s quality, which makes improving the graph quality important. Because of the importance of graph quality, this article will discuss how you can improve the quality of your graph used for machine learning.
The motivation for this article is that I am working on a project involving graphs. The quality of the graphs I create is essential to the performance of my community clustering algorithm, which is why I have spent a lot of time theorizing how the quality of the graph can be improved. I tested each idea I will mention in this article on my own graph. Some of the ideas improved the quality of my graph, some decreased the quality, and some had a neutral effect. If you want to learn more about the impact each idea can have on your graph, you can read my Towards Data Science article on testing graph quality below: