Skip to main content
Uber logo

Start ordering with Uber Eats

Order nowOrder now

Start ordering with Uber Eats

Install the appInstall the app

Applying Machine Learning in Internal Audit with Sparsely Labeled Data

March 2, 2021 / Global
Featured image for Applying Machine Learning in Internal Audit with Sparsely Labeled Data
Table 1: Confusion matrices on both training and validation data in the dual-model architecture with 80-20 split
Table 2: Confusion matrices of fully trained models in the dual-model architecture
Table 3: Classification reports on validation data of both models with 80-20 split
Table 4: Classification reports of fully trained models of both architectures *Note: This table is solely to demonstrate an apples-to-apples comparison between single-model and dual-model architectures.
Jesse He

Jesse He

Jesse is a Data Scientist and a founding member of the data science team at Uber Internal Audit. At Uber, Jesse seeks to push the boundaries of internal audit with his passion in learning and applying creative ML solutions to solve old and new problems alike. Prior to Uber, Jesse helped pioneer efficient data flow processes and create data-driven audit methodologies at EY West with the mentorship of Shan Huang and Yao Yang. Jesse obtained his Bachelor’s degree in MIS, Accounting and Finance from Purdue University, during which he also studied aeronautical engineering, aerospace management, and obtained his commercial pilot license.

Posted by Jesse He