Imbalance Treatment Results
Abstract
This repository contains the full results of the experimentation described in paper ‘Comparison of Oversampling with Thresholding for the treatment of imbalance in medical data’ (Spyridon Konstantopoulos, George Papageorgiou, Christos Tjortjis). The results are organized in a single excel file with multiple tabs as follows: > Performance metrics (Tabs 1-8): Each corresponds to one dataset of the study and contains the average values and 95% Confidence Intervals (CI) of 12 Performance metrics and the probability threshold for each imbalance treatment scenario. > F-2 Charts (Tabs 9-16): Each corresponds to one dataset of the study and contains the F-2 & threshold chart (including the relevant 95% CI) for each imbalance treatment scenario. > Chart analysis (Tab 17): Four analytics of the F-2 charts per each dataset. > Threshold comparison (Tabs 18 & 19): F-2 Threshold sweeps for datasets HCE, DIAC & CHDF in values (Tab 18) and chart (Tab 19) formats. > Results summary (Tab 20): Average F-2 and 95% CI results per each dataset.
Fields of Study & Subjects
Imbalanced Data Classification Techniques
Machine Learning and Data Classification
Artificial Intelligence in Healthcare
Publication details
- Publication date
- 1 January 2026
- Language
- EN
- Document type
- Preprint