Understanding adjustment bias and its impact on forecast accuracy – NXP



Company Name / Department


Contact Person

Ana Glaser (US), Kai Schelthoff (NL)

 Location Eindhoven, NL
Optional remote work
Travel expenses (own account or reimbursed by the company)
Housing arranged by company No
Housing expenses (how much per month, own account or subsidized by the company) Own account
Internship compensation 
Study program
ESCF community

Full member

Start date

February, 2024


Company Description

    The world leader in secure connectivity solutions for embedded applications, NXP is driving innovation in the secure connected vehicle, end-to-end security and IOT smart connected solutions markets. 

    As the world’s 5th largest chip-maker, the company has 29,000 employees in over 30 countries, with approximately $13B in annual revenue.


    Project Description

    Project description:

    The modeling team generates statistical and ML driven forecasts that are evaluated and often adjusted by the demand managers before being delivered to the supply chain organization.

    Understanding the demand managers’ adjustment patterns and their impact on forecast accuracy enables the team to recognize this behavior and mitigate it.

    Goals of the project:

    As part of this project, we would like to analyze the forecast model review and adjustment process to  identify traits that lead to model adjustment, and their resulting impact on forecast accuracy.


      This project consists of three phases:

      • Develop an understanding of the forecast review process and the various organizations involved, benchmarking it against industry best practices and literature.
      • Analyze the forecast override logs and resulting accuracy metrics to identify significant patterns in adjustment behavior and discernable accuracy outcomes.
      • Propose a solution to effectively identify predictably wrong override behaviors and develop a prototype to demonstrate its application.


      Essential student knowledge:

      Python, Machine Learning methodologies




        More information: escf@tue.nl  


        logo PostNL 240x140