Automated Machine Learning Made Real

Data scientists agree: most of the work in building your predictive model is choosing the algorithm and its hyperparameters.   Auger.AI makes this process easy and automatic.   Auger uses its industry-widest set of supervised learning algorithm implementations and its unique reinforcement learning based "smart search" to explore the problem space of available settings for them.  The result is optimal prediction effectiveness with the lowest effort.   

Manage and Optimize Your Features 

Auger takes your input data, determines the nature of each feature and easily creates derived features such as time series averages.  It analyzes errors, missing values and summarizes the distribution in your dataset.  It determines which features are adding to the effectiveness of the model.  

What's Different About Auger

• Open source, Jupyter-based

• Hosted and web accessible

• Best automated machine learning selection/optimization of algorithms & hyper-parameters.

• Auger's AutoML even manages your feature set (with deep learning based feature generation and feature pruning)

• Visual feature management including robust transformation and generation/summarization

• The most prebuilt algorithms including the first prebuilt reinforcement learning framework supporting reinforcement learning

• Affordable billing model based on hosting fees