As many who have been following this space know, the evolution of analytics has compelled many organizations to invest in predictability – that is, predicting future outcomes from existing datasets. One of the ways forward-looking organizations are seeking to resolve this is through machine learning (ML), the process of developing, testing, and applying predictive algorithms to achieve this goal. And the current reality is this is often outside of the scope of the typical “day job” of today’s developers and requires the specialization of data scientists.
There are now many tools in the market to assist data scientists in their ML activities, several of which can be very time consuming. In fact, one of the most challenging ML activities for data scientists is featurization, which is the process of converting disparate forms of data into numeric data so it can be used for basic ML algorithms.
Just take a moment to think of all the various sources and formats of data – text, images, videos, graphs, various database tables, time-series, categorical features… you can easily see the challenge this presents. The solution is to apply featurization so the conversion of these types of data into numerical features can easily feed an ML model. And it’s not surprising that featurization can often be the largest factor in the performance – and ultimately success – for many ML models. Here’s where Neo4j and AWS come in!
One of the factors that propels Neo4j as the category leader is our emphasis on innovation – we are the only graph technology to offer graph data science, a connected data analytics and machine learning platform that helps you understand the connections in big data to answer critical questions and improve predictions.
With over 65 pretuned graph algorithms and machine learning modeling to analyze your connected data, Neo4j Graph Data Science is the only connected data analysis platform that unifies the ML surface and graph database into a single workspace. And combined with AWS Sagemaker, a platform that enables developers to create, train, and deploy machine-learning models in the cloud, Neo4j and AWS provide a seamless solution for enhanced ML models through graph feature engineering.
To highlight this technology partnership, Neo4j and AWS co-authored a blog piece to demonstrate how our combined solution resolved the challenge of featurization. If you are a data scientist or vested in building successful ML models, we invite you to read our blog “Graph Feature Engineering with Neo4j and AWS Sagemaker” where you’ll learn how to:
- Deploy Neo4j Enterprise Edition
- Load a dataset in Neo4j Graph Database within the AWS SageMaker Studio
- Use Neo4j Graph Data Science to compute graph embedding on the dataset
- Run a Sagemaker job and inspect/analyze the output