Cortex is a point-and-click automated machine learning platform which allows any business access to machine learning. Cortex puts ML in the hands of everyone from marketers to data scientists to execs, providing them with a simple UI to build and operationalize predictive machine learning models.
Create predictive models with just a few clicks — no coding required. Cortex currently supports three types of ML models:
- Event Time Series Models: These models train on past events and predict whether they will occur again in a given time window
- Look Alike Models: These models train on a set of positive labels and search through an unlabeled set to find other data points with similar attributes
- Classification Models: These models train on a set of positive and negative labels in order to classify sets of unknown labels
Cortex leverages Integrated Automated Machine Learning technology (Integrated AutoML) to automate every step of the machine learning process, creating an end-to-end solution that works at the speed of your business. Cortex automatically takes in raw data from a variety of sources (Redshift, BigQuery, APIs) and formats (JSON, MRSS, CSV) and processes it into a form that can be used by its state-of-the-art learning algorithms. From real-time data ingestion to advanced model training and deployment, Cortex learns the optimal path through each step for each problem so that you can quickly and effectively optimize toward your business goals.
Automated Feature Cleaning
Cleaning data, removing outliers, accounting for missing features, normalizing input features, and merging multiple sources can be a painful and time-consuming process. Cortex learns from every deployment, and has robust and automated frameworks for transforming and managing your data.
Automated Feature Engineering
Cortex uses its experience with global Fortune 500s to transform raw data into complex features which maximize the predictive power of your models. Transformations include time quantization, sequence generation, uncommon data removal, summation of categorical events, and others.
Automated Model Selection & Learning
Cortex constantly trains, tests, and validates thousands of models, including state-of-the-art deep neural networks and gradient-boosted decision trees. Cortex evaluates the performance of each model and selects the winner on an ongoing basis. Cortex’s predictive platform adapts in real-time to ongoing data as it’s ingested, constantly learning and adapting to optimize for your strategic business goals.