Welcome to Elipsa!

Welcome to the getting started guide.  Elipsa is a no-code machine learning platform designed with ease of use and flexibility in mind. We help organizations quickly build, deploy, and scale AI projects without the cost and complications of traditional solutions.

 

Below, you will find a series of guides to help you through the process of building and deploying machine learning models in Elipsa.  

Guides to Get Started

Create Models

Elipsa automates the process of building and deploying machine learning models.  The type of model selected will depend on your use case and desired output.  Below you will find sample guides on how to create various types of models with clicks not code.

It is free to build new models in our sandbox environment. These free trial models can be accessed for one week from the date of creation.  After one week, trial models are deleted from the system. At any point during the one-week period, you have the option to convert your trial model into a paid production model.  Production models are saved to your account for ongoing use and are subject to elipsa's standard pricing.  Detailed pricing can be found HERE.

Outlier Detection

Build a model to predict abnormal events/states in your data.

Sample Use Case:

Predictive Maintance

Sample File: Turbofan Dataset

Predict a Value

Build a model to predict a numerical value

Sample Use Case:

Remaining Time Till Failure

Sample File: Turbofan Dataset

Predict an Event

Build a model to predict the likelihood of an event occuring

Sample Use Case:

Material Processing Defect Detection

Sample File: Mining Process

Understanding Model Results

Once a model is created, elipsa generated a confidence report that illustrates in simple terms how well the model performs on new data.  This allows a user to quickly and easily understand the AI model's strengths and weaknesses before putting it into production.  Elipsa's platform focuses on explainability to strip out as much data science technical jargon that we can.  The guides below will help to explain the metrics for each model type so that you can determine whether the model meets your needs.

Outlier Detection

Understand the results of Outlier Detection models

Predict a Value

Understand the results of models that predict a value

Predict an Event

Understand the results of models that predict the likelihood of an event

Make Predictions

Once models are created, users have the ability to test a model by making predictions against it with new data in our user interface. Predictions can be made on single snapshots of data or batches of multiple data points.  If the user chooses to save a model, a unique API endpoint will be created allowing the user to connect Elipsa to other applications for future predictions.

The guides below explain the process of making predictions on the Elipsa Platform as well as via our APIs.

Make Predictions on Platform

Perform batch predictions on new data directly in the Elipsa platform

Make Predictions via API

Connect Elipsa models directly to your existing applications to make predictions on steaming data via API