In recent years, the use of precise age assessment methods on living individuals has become increasingly important in the forensic field. The aim of this study was to develop, for the first time, a Bosnian and Herzegovinian-specific formula based on our local dataset, utilizing the LASSO model to uncover and leverage the unique characteristics of the Bosnian-Herzegovinian population. This research is undertaken to address the need for a Bosnia-Herzegovina-specific equation capable of accurately predicting a child’s age using local data. The dataset covering the details of 205 Bosnian boys and girls, spanning from 7 to 13 years, to maintain representativeness and reduce potential age-related biases. Three distinct models were considered, each inspired by existing approaches but adapted to the specifics of the local dataset. The first model, based on the Belgrade equation, incorporated predictors x3, x7, and N0. The second model extended the European equation by including N0, x5, s, and the interaction term s⋅N0. The third model adopted a linear regression framework with Lasso regularization, which used all available predictors (x1, x2, x3, x4, x5, x6, x7, s, N0, s⋅N0). It would be recommended that in the future the BH equation that has the form of the European one that specifically uses the x5 tooth with gender, sum and product be used. A linear model with fewer parameters is generally preferred, as the results between Lasso and such models are ultimately negligible. In this study, we explored the use of Lasso model as a method to identify which teeth contribute most to dental age estimation within a linear modeling framework. By comparing Lasso’s performance to the customized European formula for the Bosnian-Herzegovinian population, we found that the results were similar, and in some cases, Lasso even performed slightly better. Given that the European formula relies on fewer features, it presents a practical and efficient alternative as a final model for Bosnian-Herzegovinian dental age estimation.
