Erratum
With regard to the referenced publication 1, we would like to clarify and correct some formulations concerning the applied graphical probabilistic procedure.
- We would like to clarify that our paper describes a Naїve Bayes Classifier, and thus the associated language is drawn more from machine learning than it is from classical Bayesian methodology. Therefore, in the paper we use terms such as significance and testing which in fact have no particular use in the Bayesian approach.
- The statement: “The main component of this equation is LR, which is a quotient of two conditional probabilities, that is, probability of the observed multilocus genotype Mj given that a person under investigation has specific eye color Pr(Mj|Hk = x) and the sum of probabilities of the observed multilocus genotype Mj given that a person under investigation has different eye color
” on p. 882, left column, lines 1–6 should be corrected. When the purpose is to compare one proposition versus several alternative propositions, then the general form of the likelihood ratio involves also prior probabilities:
- The sentence: “Each node represents an uncertain state of a variable, and arrows between nodes represent links among these variables.” on p. 882, left column, lines 16–18 should have a meaning: “Each node represents the uncertainty associated with each possible state a variable may take.”
- The sentence: “Information put into both kinds of tables is called a priori beliefs.” on p. 882, right column, lines 1–2 should have a meaning: “Information put into both kinds of tables represent a prior beliefs.”
- The sentence: “A model with two alternative variations was developed to predict the iris color probabilities from genetic data.” on p. 882, right column, lines 7–8 should have a meaning: “A model with two alternative variations was developed to calculate the posterior probability of the iris color given genetic data.”
- The sentence: “The value of the LR representing the best discrimination between light and dark eye color calculated from our results equaled 32.3, which should be interpreted to mean that it is 32 times more probable to have dark eye color than to have light eye color, given this particular genotype.” on p. 885, left column, line 11 should have a meaning: “The value of the LR representing the best discrimination between light and dark eye color calculated from our results equaled 32.3. This means that determination of this particular genotype is 32 times more probable if an investigated subject had dark eye color than if this individual had light eye color.”
- The sentence: “This was done since we rarely have prior knowledge concerning ancestry of the identified subject, and thus, it is more reasonable to rely solely on genotypic data.” on p. 885, left column, line 29 should have a meaning: “This was done because we rarely have prior knowledge concerning ancestry of an investigated subject, and thus, it is more reasonable to rely solely on genotypic data.”