From the course: Certified Analytics Professional (CAP) Cert Prep

Unlock this course with a free trial

Join today to access over 25,300 courses taught by industry experts.

Deploying a model

Deploying a model

- [Narrator] You can run Neo4j on your own development machine and experiment with it. But when it comes to deploying your fraud detection solution, like Neo4j, to a production environment, there are several careful steps to take to ensure success. Validation is one of the first things to be done in preparation for deployment. The challenging part here is the scarcity of test datasets. Fraudsters do their best to hide their transactions and analytics specialists may not have access to a real life dataset due to security and privacy concerns. An alternative is a synthetic dataset generated by a script with intentionally embedded fraudulent transactions. The next step is developing a deployment plan. An important decision to make is location. Your Neo4j instance can reside on a local server or in the cloud. Another thing to consider is model requirements. Your stakeholders may demand a certain performance level. When you're…

Contents