Dear all,
Can anybody help me out with this? I'm looking for a dataset suitable for discriminant analysis, preferably 5 groups or less. The data are obtained with analytical chemical methods such as chromatography, infrared spectrometry etc. What I would like to do is test three different normalization methods and compare the results of Discriminant Analysis to see to what extent the normalization method influences the distinction between the different groups. Can someone provide me with such a data set?
very much obliged,
Marcel
Below are some public datasets that are suitable for discriminant analysis. These datasets vary in size and complexity, offering a range of challenges for discriminant analysis. You can further explore each dataset's details, licensing, and citation requirements:
1. UCI Machine Learning Repository https://uw.pressbooks.pub/appliedmultivariatestatistics/chapter/discriminant-analysis/
2. Mendeley Data: https://data.mendeley.com/datasets/pd6gsm2m9r/1
3. Kaggle: https://www.kaggle.com/datasets
One of the most well-known dataset for these purposes is the Anderson-Fisher Iris dataset. It is well investigated (since 1936) and everyone knows what the outcome should be. By this it is a handy set for investigating the effect of different transformations, distances, ordinations, normalisations, etc...
https://en.wikipedia.org/wiki/Iris_flower_data_set
A resource for this dataset is here: https://archive.ics.uci.edu/dataset/53/iris
However, today this dataset comes along with a couple of matrix oriented programming languages. In R it is directly implemented and can be accessed by a single (predefined) variable:
> plot (iris, col = iris$Species)
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training