News

New Enterprise Graph Framework for Data Scientists Leverages Machine Learning

• By John K. Waters
• 04/08/2020

Graph database platform provider Neo4j this week announced the availability of a new enterprise graph framework designed to help data scientists leverage largely untapped relationships and network structures that have the potential to be highly predictive.

The company is promoting its new Neo4j for Graph Data Science as the first data science environment built to harness the predictive power of relationships for enterprise deployments. 

Neo4j is the lead commercial sponsor behind the Neo4j open source NoSQL graph database implemented in Java. The company's new data science framework combines a native graph analytics workspace and graph database with scalable graph algorithms and graph visualization for "a reliable, easy-to-use experience," the company says.

The term "graph data science" can be confusing. Alicia Frame, lead product manager and data scientist at Neo4j, explained it in a recent blog post: "Simply put, graph data science is connected data science, where relationships are first-class citizens and the connections between your data points can be used to make better, more accurate predictions. Instead of thinking about your data in rows and columns, you begin to consider the importance of the relationships between those data points, which are potentially far more valuable for predictive accuracy than your flat tabular descriptors."

Graph data science uses multidisciplinary workflows that combine queries, statistics, graph algorithms and machine learning techniques to leverage the relationships and topology of connected data to artificial intelligence (AI) applications, she explained. Graph data science uses expressive Cypher queries to interrogate a graph and find local patterns. Visualization tools, such as Bloom, are sometimes used to explore, summarize and interact with the data.

The new Neo4j for Graph Data Science framework is designed to enable data scientists to operationalize better analytics and machine learning models that infer behavior based on connected data and network structures Frame described. The framework, she said in a statement announcing the product release, is intended to provide the most expeditious way to generate better predictions.

"A common misconception in data science is that more data increases accuracy and reduces false positives," she explained. "In reality, many data science models overlook the most predictive elements within data -- the connections and structures that lie within. Neo4j for Graph Data Science was conceived for this purpose -- to improve the predictive accuracy of machine learning, or answer previously unanswerable analytics questions, using the relationships inherent within existing data." 

The Neo4j for Graph Data Science framework enables data scientists to answer questions that are addressable only through understanding relationships and data structures, Frame said. Graph algorithms are a subset of data science tools that capitalize on network structure to infer meaning and make predictions such as: cluster and neighbor identification through community detection and similarity algorithms; influencer identification through centrality algorithms; and topological pattern matching through path-finding and link prediction algorithms.

The company offered the following example of how the Graph Data Science framework might be applied to the detection of identity fraud and fraud rings, a problem that spans market verticals, from financial services and insurance to the government sector and tax evasion. "Even the smallest predictive improvement translates into millions of dollars of savings," the company said.

The Neo4j for Graph Data Science framework is available now.