Sidebar

ROC regression analysis is an advanced new method for the determination or validation of diagnostic and operational criteria and characteristic thresholds. For example, ROC regression can be used to establish (or validate) criteria for healthy and sick individuals on the basis of their body temperature. If the body temperature is below 37 oC, then the person is likely to be healthy, whereas if the body temperature is above 37 oC, then the person is likely to be sick. The body temperature of 37 oC is considered as the threshold, while the body temperature variable can be regarded as a biomarker for the state of health (at least in the event of particular health conditions).

ROC curveA marker or biomarker is typically determined on the basis of a ROC curve that is a dependence of the fraction (or rate) of true-positive prediction outcomes on the fraction of false-positive prediction outcomes - like in the figure on the right. The determined threshold in the figure corresponds to the dot with the error bars indicating the statistical errors of the threshold. 

ROC regression analysis and markers/biomarkers can be used in a variety of different fields including:

  • Medical research and determination of effective diagnostic criteria for different conditions based on physiological and psychological parameters, factors and/or their combinations;
  • Ecology and environmental sciences for the development of markers and biomarkers for the determination of health and condition of species, habitat and environment;
  • For the development of reliable markers determining profitable operation and optimal operation conditions for franchisee businesses, multiple business outlets, etc. on the basis of known variables, factors or their combinations;
  • Agriculture and forestry for the determination of criteria and optimal conditions for external interventions to maximise productivity and output; 
  • Psychology and social sciences for the identification and characterisation of markers of individual and social behaviour, social health, social and political trends, etc. 

At ACHSOL, we have extensive expertise in ROC regression analyses and marker development, including in business, medical research, ecology and other areas.