Tutorial on Evaluation Indices of Statistical Performance: A Review Study

Farzan  Madadizadeh; Hooman Yekrang  Safakar; Bita  Forootani; Malihe  Bolukyazdi; Zohreh  Khosravani Shooli; Sajjad  Bahariniya

doi:10.18502/jchr.v12i4.12659

Farzan Madadizadeh Departments of biostatistics and Epidemiology, School of public health, Shahid Sadoughi University of Medical Sciences, Yazd, Iran
Hooman Yekrang Safakar Department of Nutritional Sciences, School of Public Health, Shahid Sadougi University of Medical Sciences, Yazd, Iran
Bita Forootani Department of Nutritional Sciences, School of Public Health, Shahid Sadougi University of Medical Sciences, Yazd, Iran
Malihe Bolukyazdi Department of Nutritional Sciences, School of Public Health, Shahid Sadougi University of Medical Sciences, Yazd, Iran
Zohreh Khosravani Shooli Department of Nutritional Sciences, School of Public Health, Shahid Sadougi University of Medical Sciences, Yazd, Iran
Sajjad Bahariniya Department of Health Services Management, School of Public Health, Shahid Sadoughi University of Medical Sciences, Yazd, Iran

DOI: https://doi.org/10.18502/jchr.v12i4.12659

Keywords: Data Accuracy, Likelihood Ratio Test, Sensitivity and Specificity, Statistics

Abstract

Statistical indicators are essential parts of research in many scientific fields such as health and treatment. These indicators play a major role in the evaluation of many health indicators in the general population and can help predict future issues. Statistical indicators are needed to evaluate performance of the tests. Two of the primary indicators are sensitivity and specificity, and other indices are obtained from them. In this tutorial study, evaluation indicators of statistical performance such as false negative rate (FNR), false positive rate (FPR), false discovery rate (FDR), false omission rate (FOR), bookmaker informedness (BM), markedness (MK), diagnostic odds ratio (DOR), positive likelihood ratio (PLR), negative likelihood ratio (NLR), prevalence threshold (PT), threat score (TS), prevalence (P), Fowlkes-mallows (FM), Phi-coefficient or Matthews correlation coefficient (MCC) and F1 score have been reviewed.