Oracle recently confirmed the launch of the Oracle Cloud Data Science Platform. The system is developed from Oracle Cloud Infrastructure Data Science, allowing businesses to develop, manage and deploy machine learning systems. The new technology incorporates a number of features like project sharing, team security and auditability. The product determines the most efficient training datasets through a refined algorithm and model evaluation process.
The Infrastructure Data Science includes automated data workflows and tuning that automates the procedure of testing against several algorithms and other configurations. The automated predictive features enhance the overall process by selecting key sections from larger datasets.
The model evaluation tool creates a range of metrics and visualisations to analyse overall model performance against certain datasets and allows for models to be ranked over time, enabling the best performance in production.
The new platform includes a number of new features designed to improve data science results. The Infrastructure Data Catalog enables users to manage and measure data on the vendor’s cloud. The Infrastructure Data Flow system is a completely managed big data service allowing users to operate Spark applications with no infrastructure required.
In a recent statement, Greg Pavlik, the senior VP of Production Development at Oracle Data and AI services explained that efficient machine learning models are critical in delivering successful data science projects.
Pavlik highlights that the level of success can be hindered by the volume and range of data available to businesses. Pavlik states that Oracle is improving the overall productivity of data scientists by automating workflows and providing stronger collaborative support to ensure data science projects generate real value for companies.