Efficient information theoretic strategies for classifier combination, feature extraction and performance evaluation in improving false positives and false negatives for spam e-mail filtering

2005 ◽  
Vol 18 (5-6) ◽  
pp. 799-807 ◽  
Author(s):  
V. Zorkadis ◽  
D.A. Karras ◽  
M. Panayotou
Keyword(s):  
Feature Extraction ◽  
Performance Evaluation ◽  
False Positives ◽  
Classifier Combination ◽  
False Negatives ◽  
Information Theoretic ◽  
Efficient Information ◽  
And Performance ◽  
E Mail

scholarly journals A review of feature extraction and performance evaluation in epileptic seizure detection using EEG

2020 ◽  
Vol 57 ◽  
pp. 101702 ◽  
Author(s):  
Poomipat Boonyakitanont ◽  
Apiwat Lek-uthai ◽  
Krisnachai Chomtho ◽  
Jitkomut Songsiri
Keyword(s):  
Feature Extraction ◽  
Performance Evaluation ◽  
Epileptic Seizure ◽  
Seizure Detection ◽  
Epileptic Seizure Detection ◽  
And Performance

Performance of a web-based application measuring spot quality in dried blood spot sampling

2019 ◽  
Vol 57 (12) ◽  
pp. 1846-1853 ◽  
Author(s):  
Herman Veenhof ◽  
Remco A. Koster ◽  
Randy Brinkman ◽  
Enes Senturk ◽  
Stephan J.L. Bakker ◽  
...  
Keyword(s):  
False Positives ◽  
Dried Blood Spot ◽  
Therapeutic Drug ◽  
False Negatives ◽  
Web Based ◽  
User Test ◽  
Spot Quality ◽  
System Usability ◽  
And Performance ◽  
Blood Spot

Abstract Background The dried blood spot (DBS) method allows patients and researchers to collect blood on a sampling card using a skin-prick. An important issue in the application of DBSs is that samples for therapeutic drug monitoring are frequently rejected because of poor spot quality, leading to delayed monitoring or missing data. We describe the development and performance of a web-based application (app), accessible on smartphones, tablets or desktops, capable of assessing DBS quality at the time of sampling by means of analyzing a picture of the DBS. Methods The performance of the app was compared to the judgment of experienced laboratory technicians for samples obtained in a trained and untrained setting. A robustness- and user test were performed. Results In a trained setting the app yielded an adequate decision in 90.0% of the cases with 4.1% false negatives (insufficient quality DBSs incorrectly not rejected) and 5.9% false positives (sufficient quality DBSs incorrectly rejected). In an untrained setting this was 87.4% with 5.5% false negatives and 7.1% false positives. A patient user test resulted in a system usability score of 74 out of 100 with a median time of 1 min and 45 s to use the app. Robustness testing showed a repeatability of 84%. Using the app in a trained and untrained setting improves the amount of sufficient quality samples from 80% to 95.9% and 42.2% to 87.9%, respectively. Conclusions The app can be used in trained and untrained setting to decrease the amount of insufficient quality DBS samples.

scholarly journals Information-Theoretic Analysis and Performance Evaluation of Optimal Demappers for Multi-Layer Broadcast Systems

2018 ◽  
Vol 64 (4) ◽  
pp. 781-790
Author(s):  
Eduardo Garro ◽  
Jordi Joan Gimenez ◽  
Peter Klenner ◽  
David Gomez-Barquero
Keyword(s):  
Performance Evaluation ◽  
Theoretic Analysis ◽  
Information Theoretic ◽  
And Performance
Sign in / Sign up

Export Citation Format

Share Document