Users may combine and transform datasets from multiple sources such as RDBMS, Cloud Storage, and Local, without any coding with Daria’s GUI. Daria then automatically detects variable types and anomalies and suggests appropriate data transformations. Users can apply any predefined data processing function to this process, including oversampling, scaling, and normalization, with a mere drag and drop.
Daria automatically selects the best machine learning model for each dataset from a number of possible combinations of algorithms and hyperparameters. Training results are organized onto a highly detailed analysis page in the form of intuitively designed charts and tables, allowing for a quick evaluation of model performance.
The results generated from model training can be deployed straight to production via Daria’s RESTful API. Users receive Real-time/Batch predictions, while also managing and monitoring model performance with Daria’s help.