Learning Data Science: Why a High R^2 Can Be Misleading
Explains why a high R-squared can be misleading in regression models, using examples like mean and polynomial models to illustrate overfitting.
Holger K. von Jouanne-Diedrich is a Full Professor for Information Management and CRM in Germany, AI fellow at Zukunft-Fabrik.2050, and experienced management consultant specializing in AI strategy, AI projects, and organizational transformation.
6 articles from this blog
Explains why a high R-squared can be misleading in regression models, using examples like mean and polynomial models to illustrate overfitting.
Explores In-Context Learning (ICL) for tabular data using the TabPFN R package, a transformer-based foundation model that predicts patterns without retraining.
A tutorial on building a transformer-based language model in R from scratch, covering tokenization, self-attention, and text generation.
Explains the difference between weather and climate using chaos theory and the Lorenz system, with R code examples.
A simulation using multi-agent models in R to replicate complex stock market behavior through technical and fundamental trading strategies.
A quick-start guide to using the R programming language for data analysis, covering installation, data exploration, and basic plotting with the iris dataset.