Embedded Linux Quality Assurance: How to Not Lie with Statistics

Explore statistical quality assurance techniques for embedded Linux systems in this 53-minute conference talk by Wolfgang Mauerer of Siemens AG/OTH Regensburg. Learn best practices for designing and running automated statistical tests, evaluating data, and avoiding common mistakes in interpreting results. Gain practical insights using real-world examples and the open-source R language. Discover how to measure and analyze data effectively, understand different types of data, and apply nonparametric methods. Delve into topics such as histograms, transformations, summary statistics, cumulative distributions, and linear regression. Enhance your ability to ensure reliability, error rates, and maximal latencies in embedded Linux applications, including real-time and safety-critical systems.

Introduction
How to Not Lie with Statistics
Testing
Certification
Why Statistics
Dealing with Data
Measuring Data
Tidy Data
Data in messy form
A proper statistical environment
How to understand data
Types of data
Histogram
Transformations
Nonparametric method
Summary statistics
Summary statistics example
Comparing data sets
Cumulative distributions
Common mistakes
Linear Regression
Linear Regression Results
But Fix Commits
Summary