Gretel.ai: Open-Source Artificial Intelligence Tool To Generate New Synthetic Data

Abstract

Nowadays, machine learning is widely employed to solve real-world problems, particularly in medical and diagnostics fields. However, to trained a machine learning model required a massive amount of data making it challenging due to the demand access of medical data is rigid since the data is kept confidential, secure, and difficult to obtain. Therefore, this paper introduces Gretel.ai, an artificial intelligence tool for generating synthetic data that related to Autism Spectrum Disorder (ASD). The results demonstrate the proposed framework for generating synthetic data using the Autism Quotient 10 (AQ10) screening instrument growth from 1054 original records to 5000 synthetic records while preserving the primary characteristics of the originals dataset’s features. Additionally, the Gretel.ai also provide a quick synthetic report that quantifies the utility on exploratory data analysis by presenting data summary statistics comparing the synthesized and the original training data comparisons.