Machine learning-based clustering model to create auditable entities

US11630852B1 · kind B1 · utility

Assignee

• Wells Fargo Bank, N.A. · US

Inventors

• Tylor Christopher Mondloch · San Antonio, US
• Richard Penfil, II · San Antonio, US
• Brian Karp · San Francisco, US
• Corinne Elizabeth Larsen · Robbinsdale, US
• Jeremy Salden · San Francisco, US
• Kelley Impoco · Charlotte, US
• Jeremy Paynes · Lakeville, US

Key dates

• Filing date: Jan 8, 2021
• Grant date: Apr 18, 2023
• Expiry date: Mar 23, 2041

Classification

• Technology area (CPC G): Physics
• CPC primary: G06Q50/26
• WIPO field: IT methods for management
• WIPO sector: Electrical engineering

Abstract

Techniques are described for automatic creation of optimal auditable entities (AEs) using a machine learning (ML)-based clustering model. The clustering model, when executed on one or more computing devices within an audit system of a company, is configured to automatically cluster the company's business processes into AEs based on similarity analyses of business process attributes. More specifically, in some examples, the clustering model ingests business processes and their corresponding attributes from a database, automatically clusters together business processes to achieve maximum intra-cluster similarity scores, and outputs the final clusters as model AEs. The resulting model AEs may be used as functional units for internal audits of the company's business processes. The resulting model AEs may improve audit efficiency due to the model AEs including only highly similar business processes. In addition, the resulting model AEs may enable more accurate assignment of audits based upon auditor experience and technical skills.