Statistical data analysis and interpretation of a small dataset (23 samples)

Statistical data analysis of a small dataset containing the performance results (accuracy, precision, recall, f1-score, AUC score) obtained by different configurations of neural networks architectures.



Statistical data analysis of a small dataset containing the performance results (accuracy, precision, recall, f1-score, AUC score) obtained by different configurations of neural networks architectures.

The dataset contains 23 lines of results obtained by the 23 configuration tested.
I expect full statistical analysis of your results including descriptive analysis and statistical tests.


The analysis of the results will be attached to a journal article and should be done as professionally as possible. It doesn't have to be too long, as the dataset with the performance metrics obtained is very small, but should look like a work done by someone with advanced skills in statistics.

I was thinking about a descriptive analysis + interpretation of the performance metrics in order to find the best configuration of each architecture
and applying the appropriate statistical tests to prove the configuration used for the ensemble model from BuiltCNN, VGGNet, ResNet, InceptionResNet (in bold) is the best ones (at different confidence levels).
However, as the dataset is so small, I lack the skills to do this task.

I ordered the results in an excel spreadsheet containing 23 lines with the results obtained. The metrics resulted from a binary classification using deep neural network of images for an object detection task in aerial imagery.


Deliverables:
Word document describing the statistical analysis conducted. The document should be written in a technical manner.