Dr. Tamara Schamberger
Data Science Group, Faculty of Business Administration and Economics, Bielefeld University

Structural Equation Modeling with Composites

Structural equation modeling (SEM) is a widely applied method in various disciplines, including business, economics, psychology, sociology, and medicine. Its popularity is arguably due to its ability to model and assess theories comprising theoretical concepts and to account for various forms of measurement errors.

Traditionally, SEM has been primarily used to model theoretical concepts as latent variables, i.e., variables that explain the variance-covariance structure of their related variables. In recent years, the composite, i.e., a linear combination of more elementary variables that transmits all information between the variables that make up the composite and all other variables in the model, has gained increasing attention as a second type of theoretical construct in SEM. However, due to the types of relations in a composite model, the integration of composites in a structural equation model is not straightforward.

This presentation will introduce the composite model and discuss various model specifications for incorporating composites into a structural equation model. Finally, the presentation will discuss the main approaches to estimating composite models in SEM.

20.06.2024

Dr. Max Westphal
Fraunhofer-Institut für Digitale Medizin, Bremen

Improving model and algorithm evaluation in predictive modelling (one step at a time)

In applied predictive modelling, practitioners usually face many methodological questions which are related to different objectives of the overall development and evaluation pipeline, e.g.:

“How should I split my data for training, comparison and testing purposes?”
“Which metric is suitable for assessing the utility of the developed models?”
“How can I quantify uncertainty of my evaluation results?”

While “standard” answers to these questions exist, it is rarely clear if and in what sense they are optimal or even admissible for the prediction task at hand. In the first part of the talk, we will consider three relevant and concrete examples of such questions related to model evaluation. Furthermore, we will illustrate how the derivation of more appropriate (than “standard”) answers is in principle possible based on recent methodological work.

However, finding problem-tailored solutions in the scientific literature can also be difficult and time-consuming, in particular for practitioners without a specific data science background. In the second part of the talk, I will therefore introduce the expert system mlguide which is currently under active development at Fraunhofer MEVIS. mlguide aims to provide interactive support to practitioners for solving methodological questions in applied machine learning problems, such as those mentioned before. I will illustrate the overall system architecture, some insights into the guidance engine, and current limitations. To conclude, I will give an outlook on planned features and user research.

08.05.2024