Frequently asked questions

Predictive analytics you can trust

Here are the answers to some of the most frequently asked questions

How does Pecan produce business predictions?

Pecan has developed a state-of-the art deep learning platform that identifies and extracts those nuanced hidden patterns in the historical data, transforming them into meaningful data and actionable insights.

Do I need a Data Scientist to use Pecan?

Absolutely not! Pecan uses its proprietary AI algorithms to emulate the processes and services performed by experienced data scientists. With Pecan you can construct predictive models from A-Z, without any prior experience.

What types of predictions does Pecan provide?

Pecan can provide predictions for any field and use-case in predictive analytics. For example, churn prediction, lifetime value estimation, fraud detection, and lead scoring.

Do I need a data scientist for the data modeling phase?

Absolutely not! Pecan has developed a first of its kind user interface  that enables you to ingest your data and construct a model, using a simple drag-n-drop interface alone.

How do I know which data to include in the AI schema?

You don’t need to know which data to include. On the contrary, we recommend adding as much as you can. Pecan uses advanced methods and AI algorithms to determine the relevancy of the data.

How long does it take to construct a deep-learning predictive analytics model with Pecan?

Usually, it takes no longer than a few hours, and that’s for a novice user!

How do I know that the AI model works?

We rely on stringent backtesting that prevents overfitting and increase generalization.

What types of data does Pecan Support?

We merge data from various sources, such as DBs, ERP, CRM, Excel sheets and BI software products.

What databases does Pecan connect to?

Pecan connects to the following databases: SQL server, Oracle, MySQL, PostgreSQL, Teradata, Redshift, and BigQuery.

How much data do I need?

We need to analyze at least 3,000 unique entries to ensure accurate predictions.

How is deep learning different from regular machine learning algorithms used by competing solutions?

Deep learning (neural networks) is the only AI algorithm used by universal Turing machines. It can approximate any mathematical function, so if there is a regularity in the data, even the most complex and delicate one, the network will eventually identify it.

Which deep learning techniques and frameworks does Pecan use?

Pecan uses a unique and proprietary mixture of the most advanced deep-learning techniques.

How can I simulate different future outcomes, and their likelihood, using Pecan?

Our Ph.D. scientists in Pecan have developed a unique algorithm that specializes in AI counterfactual (causal) reasoning. [This option will be available soon].

How do I Install and implement Pecan in my organization?

This depends if you use Pecan as SaaS or keep everything locally.

SaaS platform
Connect your database to Pecan’s platform (via an SSH tunnel), and start creating deep learning models.

Self-hosted solution (on your virtual private cloud; VPC)
Let Pecan’s SaaS platform run on your private cloud and never let your data leave your domain.

Do I need DBA/IT engineers to deploy Pecan?

In both the SaaS and self-hosted installations and deployment, you do not need to write any code or make any infrastructure or database changes.

In low-latency high-performance (“online”) predictions, we recommend that your DB/IT engineers replicate your AI-relevant tables.

Where is my data stored and how safe is it?

In Pecan SaaS, your data is stored on Pecan’s AWS infrastructure.

In a self-hosted VPC installation, your data never leaves your own environment.