AutoML and Predictive Analytics Platform

Actionable predictions in the shortest time possible

Drive impactful, scalable business value from AI

Pecan was designed to drive business value from AI. In one intuitive platform, analysts and business people can add data, deploy off-the-shelf predictive models, and see outputs—finally bringing AutoML within reach. Learn how the platform works below.

Automated Data Preparation

Manual data prep can be the biggest time sink in predictive, adding weeks or months to a project’s timeline. Plug in your data sources and the Pecan platform automates data restructuring, encoding and cleansing, imputation, and features selection and engineering.

Modeling and Enrichment

Pecan’s proprietary AI algorithms optimize and train your predictive models to get to effective results as quickly as scientifically possible. But drift and leakage kill models, so our platform continuously monitors and optimizes your models, getting more accurate and more useful over time. And to get to results faster, Pecan will enrich your models with relevant external data.

Dashboards and Data for Action

Actionability begins from step one by design. The Pecan platform is organized into projects and models so you can stay laser-focused on your business KPIs no matter what type of model you approach them with. When it comes time to take action, our predictions connect at the data source so they’re immediately available to act on.

Speed, usability, and advanced AI

Advanced features to ensure accurate, actionable predictions

1 Data Set Maker: With Pecan you can naturally connect to a variety of databases via code-free integration—and no database manipulation. Once you’re connected, use the simple drag-n-drop interface to create ML-ready data sets.
2 Data Enrichment: Supercharge your easily-made models using data enrichment, which draws relevant global data to your predictive analysis in order to arrive at more relevant results—faster.
3 AutoPrep: Automating feature engineering, feature selection, and encoding not only saves time on one of the most complicated parts of the predictive process, but it enables your outputs to improve as Pecan constantly trains and optimizes your models, prioritizing features as they change in importance over time.
4 AutoML: Matching the nature of your data, its size, and the predictions you require, Pecan constructs a large number of deep neural networks. After an exhaustive set of recursive competitions between multiple networks, only one fully trained neural network is left – uniquely evolved and refined for optimal performance and accuracy.
5 Monitoring: Your first time using Pecan will yield predictions in just days. But it only gets better—Pecan’s AI-powered optimization will continually analyze your data to yield better, more accurate predictions as time goes on. Use the live dashboard to decide on actions and monitor performance.