Applications of statistical causal inference in software engineering

  title={Applications of statistical causal inference in software engineering},
  author={Julien Siebert},

Towards Causal Analysis of Empirical Software Engineering Data: The Impact of Programming Languages on Coding Competitions

Novel techniques that support analyzing purely observational data for causal relations are discussed, which can help answer the salient research questions more precisely and more robustly than with just purely statistical techniques.

Dataflow graphs as complete causal graphs

This paper considers an alternative approach to software design, flow-based programming (FBP), and draws the attention of the community to the connection between dataflow graphs produced by FBP and structural causal models.



Causal Inference for Theory Building in Software Evolution

This work proposes a new methodology to construct and validate theories of software evolution that utilises causal inference to define and test causal relations.

Causal Modeling, Discovery, & Inference for Software Engineering

This paper applies causal discovery techniques to a set of observational data on open source projects, and uses the results to determine some consequences of architectural flaws, and shows how causal inference may be applied to software engineering data in the future.

NUMFL: Localizing Faults in Numerical Software Using a Value-Based Causal Model

An evaluation of NUMFL with components from four Java numerical libraries, in which it was compared to five alternative statistical fault localization metrics, indicates that NUMFL is the most effective technique overall.

A general framework for blaming in component-based systems

Causal inference for statistical fault localization

This methodology combines statistical techniques for counterfactual inference with causal graphical models to obtain causal-effect estimates that are not subject to severe confounding bias and empirical results demonstrate that the model significantly improves the effectiveness of fault localization.

Bayesian causal inference in automotive software engineering and online evaluation

This study develops the BOAT framework, which enables online software evaluation in the automotive domain without the need of a fully randomised experiment, and relates the causal assumption theories to their implications in practise to provide a comprehensive guide on how to apply the causal models in automotive software engineering.

Evaluation of Causal Inference Techniques for AIOps

Preliminary results indicate that event models yield causal graphs that have high precision and recall in comparison to regression and independence testing based Granger methods.

Preliminary Causal Discovery Results with Software Effort Estimation Data

This paper will introduce causal discovery to software engineering research, and its use in the future may impact how software effort models are built.

Detecting Causal Structure on Cloud Application Microservices Using Granger Causality Models

An extensive comparative study is conducted to show the performance of the state-of-the-art linear and nonlinear Granger causality methods on both synthetic data and real-world log data from a publicly available benchmark microservice system.

From Verification to Causality-based Explications

This paper surveys approaches to formally explicate the observable behavior of reactive systems and describes how Halpern and Pearl’s notion of actual causation inspired verificationoriented studies of cause-effect relationships in the evolution of a system.