Sex Estimation Models Research Articles

Sex estimation from patellar measurements in a contemporary Italian population: a machine learning approach.

Biological sex estimation in forensic anthropology is a crucial topic, and the patella has shown promise in this regard due to its sexual dimorphism. This study uses 12 machine learning models for sex estimation based on three patellar measurements (maximum height, breadth, and thickness). Data was collected from 180 skeletons of a contemporary Italian population (83 males and 97 females) as well as from an independent sample of 21 forensic cases (13 males and 8 females). Statistical analyses indicated that each of the variables exhibited significant sexual dimorphism. To predict biological sex, the classifiers were built using 70% of a reference sample, then tested on the remaining 30% of the original sample and then tested again on the independent sample. The different classifiers generated accuracies varied between 0.85 and 0.91 on the reference sample and between 0.71 and 0.95 for the validation sample. SVM classifier stood out with the highest accuracy and seemed the best model for our study.This study contributes to the growing application of machine learning in forensic anthropology by being the first to apply such techniques to patellar measurements in an Italian population. It aims to enhance the accuracy and efficiency of biological sex estimation from the patella, building on promising results observed with other skeletal elements.

Estimation of sex from femoral bone using radiological imaging methods in Turkish population.

Sex estimation is leading to determine the biological profile in forensic medicine. The aim of this study is to research the effectiveness of logistic regression (LogR) and discriminant function analysis (DFA) to create sex estimation models on femur images obtained with Computed Tomography (CT) angiography and to address the differences of femur, which show sexual dimorphism, among populations. All parameters were measured on three planes by adjusting the 300 CT angiography images from 150 women and 150 men that focused on the left femur to the orthogonal plane with standard magnification. The subgroup, which included 30 images randomly generated from these images, was measured twice with an interval of 3 weeks by the first radiologist and once by the second radiologist. According to the Fisher's Linear Discriminant analysis, which was evaluated with ten parameters in the study, it was concluded that the power of discriminating women was 96.7%, the power of discriminating men was 98.7%, and the total discrimination power was 97.7%; these results were 98%, 99.3%, and 98.7%, respectively according to LogR. In this study, DFA and LogR analysis showed that femur provided a very good rate of sexual dimorphism. A database belonging to the Turkish population was created for the femur, allowing for comparison between populations.

Sex estimation using post-mortem computed tomographic images of the clavicle in a Malaysian population

Forensic practitioners need contemporary anthropological data for the identification of human remains. The clavicle possesses a high degree of variability in its anatomical, biomechanical, and morphological features that are sex-dependent albeit population specific. The aim of this study was to develop sex estimation models for Malaysian individuals using post-mortem computed tomographic images of the clavicle. Sample comprised scans of 2.0 mm resolution of 405 individuals (209 male; 196 female) aged between 19 to 88 years. These scans were reconstructed and visualized using Infinitt. Six clavicular measurements (i.e. maximum length, C1; midshaft circumference, C2; midshaft maximum diameter, C3; midshaft minimum diameter, C4; maximum breadth of the sternal end, C5; and maximum breadth of the acromial articular surface, C6) were obtained from these images. Data were analyzed using descriptive statistics and discriminant function analysis. Measurements taken from the images were highly precise (ICC = 0.770–0.999). There is a significant difference between all parameters and sex (p < 0.001), however none for age and ethnic group. A multivariate sex estimation model was developed: Sex = (C1*0.86) + (C2*0.236) + (C3*-0.145) + (C5*- 0.074) − 17.618; with an accuracy rate of 89.1 % and sex bias of −3.2 %. Lower accuracy rates were obtained for single variable models (61.5–83.2 %). The resultant sex discriminant models can be used for estimating sex based on the clavicle in our local forensic practice.

Reliability of predictive models based on the mesiodistal and buccolingual diameters of permanent canine teeth for sex estimation in forensic contexts: A systematic review and meta-analysis

This study aimed to assess the reliability of predictive models for sex estimation based on permanent canine size. A systematic literature review was performed by following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA). Six electronic databases were searched as the primary source of information. As a secondary source of information, a manual search was performed to identify additional relevant studies not captured in the initial search. After assessing the methodological quality and risk of bias with the Joanna Briggs Institute Critical Appraisal Tools for Systematic Reviews, the data were subjected to statistical tests for a meta-analysis of diagnostic test accuracy and Higgin’s I2 statistic to evaluate the heterogeneity between the eligible studies. The systematic search resulted in 21 studies for qualitative synthesis, and 13 of them were selected for quantitative analysis. The analysis of 25 univariate predictive models showed an estimated sensitivity of 77.2 % and specificity of 67.1 %. Meta-regression analyses were performed for dental arch, the type of diameter and dental region outcomes for these univariate predictive models. Dental arch (p = 0.029) and the dental region of measurement (p = 0.001) were significant modifiers. The analysis of 25 multivariate predictive models showed an estimated sensitivity of 82.6 % and specificity of 70.1 %. There were significant methodological limitations and substantial heterogeneity among the included studies. Based on the results, there is insufficient high-quality scientific evidence to support the safe use of predictive models based on permanent canine measurements as the exclusive method for sex estimation in forensic settings.

Different machine learning methods based on maxillary sinus in sex estimation for northwestern Chinese Han population.

Sex estimation is a critical aspect of forensic expertise. Some special anatomical structures, such as the maxillary sinus, can still maintain integrity in harsh environmental conditions and may be served as a basis for sex estimation. Due to the complex nature of sex estimation, several studies have been conducted using different machine learning algorithms to improve the accuracy of sex prediction from anatomical measurements. In this study, linear data of the maxillary sinus in the population of northwest China by using Cone-Beam Computed Tomography (CBCT) were collected and utilized to develop logistic, K-Nearest Neighbor (KNN), Support Vector Machine (SVM) and random forest (RF) models for sex estimation with R 4.3.1. CBCT images from 477 samples of Han population (75 males and 81 females, aged 5-17 years; 162 males and 159 females, aged 18-72) were used to establish and verify the model. Length (MSL), width (MSW), height (MSH) of both the left and right maxillary sinuses and distance of lateral wall between two maxillary sinuses (distance) were measured. 80% of the data were randomly picked as the training set and others were testing set. Besides, these samples were grouped by age bracket and fitted models as an attempt. Overall, the accuracy of the sex estimation for individuals over 18 years old on the testing set was 77.78%, with a slightly higher accuracy rate for males at 78.12% compared to females at 77.42%. However, accuracy of sex estimation for individuals under 18 was challenging. In comparison to logistic, KNN and SVM, RF exhibited higher accuracy rates. Moreover, incorporating age as a variable improved the accuracy of sex estimation, particularly in the 18-27 age group, where the accuracy rate increased to 88.46%. Meanwhile, all variables showed a linear correlation with age. The linear measurements of the maxillary sinus could be a valuable tool for sex estimation in individuals aged 18 and over. A robust RF model has been developed for sex estimation within the Han population residing in the northwestern region of China. The accuracy of sex estimation could be higher when age is used as a predictive variable.

Assessing the reliability and accuracy of sex estimation models utilizing sternal morphometry derived from computed tomography in the Ghanaian population

Sex estimation models are specific to populations and cannot be generalized due to genetic and environmental variabilities. This retrospective cross-sectional study, conducted between January and September 2023 at the Tamale Teaching Hospital, included 119 (50.9 %) females and 115 (49.1 %) males aged 23–82 years. Measurements, including manubrium length (M), manubrium width (MW), sternal body length (B), combined manubrium and sternal body lengths (CL), corpus sterni width at first sternebrae (CSWS1), and corpus sterni width at third sternebrae (CSWS3), were obtained from Computerized Tomographic (CT) images of the sternum using DICOM Viewer, accurate to 0.1 cm. Subsequently, sternal area (SA) and sternal index (SI) were calculated. Univariable and stepwise multivariable discriminant function analysis (DFA) and logistic regression (LR) models were developed using a training sample (70 %), and cross-validation was performed on a holdout sample (30 %). Results showed that the linear measurements, excluding M, and sternal area were higher in males than females (P<0.001), while the sternal index was <50 % in males but >50 % in females (P<0.001). Univariable sex estimation accuracies, in cross-validation, ranged from 43.7 % to 92.9 % in DFA and 50.0–92.9 % in LR. For multivariable models, the accuracy ranges were 92.9–94.3 % in DFA and 91.6–93.0 % in LR. The sternal body length was the most accurate at 90.1 % in DFA and 90.2 % in LR, with lower sex bias (male-female) in LR than in DFA (-0.1 vs. 8.7). The sternum proves valuable for sex estimation, with sternal body length as the most accurate linear measurement. However, multivariable models, particularly LR, demonstrate higher accuracy compared to DFA.

Logistic Regression Models for Sex Estimation Using Percutaneous Upper Arm and Forearm Anthropometric Dimensions in Adolescent School Children in Southwest Nigeria

Logistic Regression Models for Sex Estimation Using Percutaneous Upper Arm and Forearm Anthropometric Dimensions in Adolescent School Children in Southwest Nigeria

Sex estimation based on mandibular measurements.

Medical imaging and machine learning are beneficial approaches in physical and forensic anthropology. They are particularly useful for the development of models for sex identification based on bone remains. The present study uses machine learning algorithms to create models for sex estimation based on mandibular measurements. The sample included head CT scans of 239 adult Bulgarians (116 males and 123 females). Three-dimensional coordinates of 45 landmarks of the mandible were acquired from segmented polygonal models of the skulls of these individuals. Two datasets of mandibular measurements were assembled. The first dataset included 51 measurements: linear, projective, and angular measurements. The second dataset included 990 interlandmark distances. Seven machine learning algorithms (Support Vector Machines, Neural Network, Naïve Bayes, Random Forest, J48, JRip, and Logistic Regression) were applied to the two datasets, and the classification accuracy was evaluated by 10x5-cross-validation. The selection of the best subsets of attributes specific to each of the abovementioned algorithms was done based on the attribute importance evaluated by an attribute selection scheme. In general, the sub-symbolic algorithms achieved higher results than the symbolic ones, except for the logistic regression. The best classification model was learnt by the Support Vector Machines algorithm, which achieved an accuracy of 95.3% on a dataset described by 19 interlandmark distances. In both datasets, the application of advanced attribute selection has led to an increase in the classification accuracy of all algorithms used in the experiments.

Testing the accuracy of the SexEst software for sex estimation in a modern Greek sample

AbstractSex estimation from human skeletal remains is fundamental in osteoarcheology and forensic anthropology. The increasing availability of reference skeletal collections across the world has allowed the development of morphological and metric methods for skeletal sex estimation, some of which may be implemented in specialized computer software. The present study aims to evaluate the freely available SexEst software, which utilizes cranial and postcranial measurements, and different classification models for sex estimation, on a contemporary Greek population comprising of 227 (126 males and 101 females) adult individuals. After the calculation of intra‐observer error to assess the repeatability of the measurements, the proposed variables were tested for classification accuracy individually and in different combinations. Based on the results, the postcranial models outperformed the cranial ones in all cases and can be adequately applied on a Greek population sample. The light gradient boosting (LGB) algorithm yielded the highest correct classification rates when no missing values exist, while the linear discriminant analysis (LDA) models should only be used when dealing with missing data. The highest classification accuracy for a 0.65 posterior probability threshold was reached when utilizing a combination of postcranial variables (89.67%), while the lowest was achieved with the cranial measurement “Glabella‐occipital length” (45.00%). The same models yielded the highest and lowest accuracy for a 0.5 probability threshold, with values of 92.96% and 67.73%, respectively. Combining variables yielded higher accuracies in both skeletal regions, suggesting that the software would be more helpful in cases of intact skeletons. The loss of classification accuracy due to population specificity further corroborates the need to include different ancestries in sex estimation software.

Testing of morphological sex estimation traits with a sex‐known collection: Ottoman period skulls

AbstractSexual dimorphism patterns vary across geographic regions due to the influence of genetic characteristics and environmental factors. Therefore, sex estimation models are being developed specifically for each population or group. The applicability of morphological sex estimation methods has not been tested in Turkey. Hence, by using skulls, the present study aims to analyze the reliability of the visual morphological method and test the equations developed in different populations. The study material consists of 192 skulls (96 male, 96 female) with known sexes, excavated from Istanbul's Karacaahmet cemetery in 1925. In the present study, glabella, mastoid process, supraorbital margin, and nuchal crest traits were scored on a scale of 1 to 5 according to the instructions provided in standard protocols. Intra‐observer and inter‐observer agreements were analyzed by two experts having the same level of experience. When equations derived from other populations were applied to our samples, they exhibited high sex biases (up to 50%). Therefore, new equations were derived through binary logistic regression analysis. Glabella had the highest performance in terms of repeatability (0.83) and reproducibility (0.74), whereas the nuchal crest showed the lowest performance (0.60–0.52). The most significant sexual dimorphism was observed in the glabella. Based on cross‐validated results using a single criterion, it accurately classified 80% of females and 84% of males. The nuchal crest was not significantly affecting the sex discriminative equations (p &gt; 0.05). Multivariate equations achieved an accuracy of over 90% and cross‐validated results ranged between 80% and 90%. The results obtained from present study support the hypothesis that sexual dimorphism patterns vary under different conditions and highlight the importance of population variation in sex estimation. The models derived from the present study were found to be suitable for sex estimation from skulls and demonstrated high performance.

The impact of age on sex bias in models for height and sex estimation based on hand and foot dimensions the impact of age on sex bias

There are sex differences in age-related bone modifications after puberty. Androgens stimulate radial bone expansion in males, while estrogen stimulates endosteal apposition but limits periosteal expansion in females. The potential effect of age on observed sex biases in height and sex estimation models is most significantly relevant for forensic or bioarchaeological research that relies, at least in part, on hand and foot bone measurements of living or skeletal remains for purposes of identification or demographic reconstruction. This study sought to determine whether age affects sex biases in models for height and sex estimation which are based on hand and foot dimensions. The study was cross-sectional between January and June 2021 at the University for Development Studies. The study included 379 participants (male = 161 and female = 218) between 20 and 29 years. The hand length (HL), hand width (HW), foot length (FL) and foot width (FW) were measured twice from the left side using computer-assisted analysis. Univariable and multivariable discriminant and linear regression models were formulated for sex and height estimation respectively. Females were better classified than males with sex biases (male-female) ranging from −1.3% to −22.6% for all models. Models for height estimation were more precise in males (bias: 0.0–0.3 cm) than in females (bias: 0.3–1.4 cm). However, age did not have an impact on the observed sex biases. Height and sex estimation from foot and height dimensions may not need adjustment for age. This may, however, be limited to a given population and age group.

Sex estimation from long bones: a machine learning approach.

Sex estimation from skeletal remains is one of the crucial issues in forensic anthropology. Long bones can be a valid alternative to skeletal remains for sex estimation when more dimorphic bones are absent or degraded, preventing any estimation from the first intention methods. The purpose of this study was to generate and compare classification models for sex estimation based on combined measurement of long bones using machine learning classifiers. Eighteen measurements from four long bones (radius, humerus, femur, and tibia) were taken from a total of 2141 individuals. Five machine learning methods were employed to predict the sex: a linear discriminant analysis (LDA), penalized logistic regression (PLR), random forest (RF), support vector machine (SVM), and artificial neural network (ANN). The different classification algorithms using all bones generated highly accuracy models with cross-validation, ranging from 90 to 92% on the validation sample. The classification with isolated bones ranked between 83.3 and 90.3% on the validation sample. In both cases, random forest stands out with the highest accuracy and seems to be the best model for our investigation. This study upholds the value of combined long bones for sex estimation and provides models that can be applied with high accuracy to different populations.

Models for estimating age and sex from variables of the proximal femur in a Ghanaian population

Age and sex estimation models from the proximal femur are population-specific due to genetic and environmental variabilities. Extrapolating a femur-based age or sex estimating model from one population to another will be problematic. Proximal femur-specific age and sex estimation models are limited in Ghana. The study aimed to estimate the age and sex of an adult Ghanaian population using osteometric measurements from the proximal femur. The study was cross-sectional from January to June 2019 at the Korle-Bu Teaching Hospital (KBTH). There were 125 (male=51, female=74) participants, aged from 31 to 82 years. The head diameter-left (HDL), neck diameter-left (NDL), neck-shaft angle-left (NSAL) and the Hip axis length-left (HALL) were measured twice using a standardized radiographic technique. Discriminant and logistic regression models were formulated for sex estimation while linear regression models were formulated for age estimation and these models were then tested for reliability. Males had longer HDL and NDL than females (P < 0.050). The average sex estimation accuracy in the discriminant analysis ranged from 58.4% to 64.0% (in the original sample) and 56.8% to 62.4% (in the cross-validation sample) while in the logistic regression analysis it ranged from 58.4% to 64.0%. The HDL was better than the NDL in sex attribution but only marginal. The multivariate model (HDL+NDL) marginally improved sex estimation accuracy (64.0%) over the univariate models for HDL (61.6%) and NDL (60.0%). In general, females were better classified than males. There was no significant difference between the chronological age and the estimated age of males using HALL although the confidence interval (95%CI) was wider than expected [Bias: 1.133 (95%CI: −25.280 to 27.540). The femoral head and neck diameters or their combination are poor attributors of sex on average. Also, male adult age estimation using HALL is less precise. The use of these models in the Ghanaian population is not advised

Sex estimation from the percutaneous lengths of the femur and the ulna in a Ghanaian population using discriminant function analysis

Sex estimation models form a vital part in Forensic human identification but they are usually population-specific. This study aimed to develop and test sex estimation models for a Ghanaian population using percutaneous lengths of the femur (FL) and ulna (UL). The study was cross-sectional from June to July 2020, involving 99 adults (male: 52, females: 47), aged between 19 and 31 years. The lengths of the femur and ulna were measured using standard anthropometric techniques. All measurements were taken twice from the left side and then averaged. The sample was randomly divided into training (n = 60) and holdout (n = 39) samples before been analysed using discriminant function analysis (DFA). Cross-population studies were performed to test the reliability of the models. Males had longer femur and ulna than females (p < 0.001). Sex estimation accuracies from all the models ranged from 68.2% to 81.8% for males and 52.9% to 86.7% for females. The standardized mean difference (SMD: Cohen’s d) by sample type ranged from −0.19 to 3.08 (living samples), 0.19 to 4.73 (cadaveric samples) and 0.30 to 5.46 (skeletal samples). The SMD by population type were: Africa, excluding Mixed or White ethnicities (d= −0.02 to 3.08), Asia (d = 0.83 to 4.85) and Europe or the Americas (d = 0.30 to 3.38). When other population-specific models were tested on the holdout sample, the difference in the average sex estimation accuracy ranged from 0 to 25.6%. Sex estimation models from the lengths of the femur and ulna are specific to a the studied population and the type of sample used.

Use of deep learning in forensic sex estimation of virtual pelvic models from the Han population

Accurate sex estimation is crucial to determine the identity of human skeletal remains effectively. Here, we developed convolutional neural network (CNN) models for sex estimation on virtual hemi-pelvic regions, including the ventral pubis (VP), dorsal pubis (DP), greater sciatic notch (GSN), pelvic inlet (PI), ischium, and acetabulum from the Han population and compared these models with two experienced forensic anthropologists using morphological methods. A Computed Tomography (CT) dataset of 862 individuals was divided into the subgroups of training, validation, and testing, respectively. The CT-based virtual hemi-pelvises from the training and validation groups were used to calibrate sex estimation models; and then a testing dataset was used to evaluate the performance of the trained models and two human experts on the sex estimation of specific pelvic regions in terms of overall accuracy, sensitivity, specificity, F1 score, and receiver operating characteristic (ROC) curve. Except for the ischium and acetabulum, the CNN models trained with the VP, DP, GSN, and PI images achieved excellent results with all the prediction metrics over 0.9. All accuracies were superior to those of the two forensic anthropologists in the independent testing. Notably, the heatmap results confirmed that the trained CNN models were focused on traditional sexual anatomic traits for sex classification. This study demonstrates the potential of AI techniques based on the radiological dataset in sex estimation of virtual pelvic models. The excellent sex estimation performance obtained by the CNN models indicates that this method is valuable to proceed with in prospective forensic trials. Key points Deep learning can be a promising alternative for sex estimation based on the pelvis in forensic anthropology. The deep learning convolutional neural network models outperformed two forensic anthropologists using classical morphological methods. The heatmaps indicated that the most known sex-related anatomic traits contributed to correct sex determination.

Machine learning approaches for sex estimation using cranial measurements.

The aim of the present study is to apply support vector machines (SVM) and artificial neural network (ANN) as sex classifiers and to generate useful classification models for sex estimation based on cranial measurements. Besides, the performance of the generated sub-symbolic machine learning models is compared with models developed through logistic regression (LR). The study was carried out on computed tomography images of 393 Bulgarian adults (169 males and 224 females). The three-dimensional coordinates of 47 landmarks were acquired and used for calculation of the cranial measurements. A total of 64 measurements (linear distances, angles, triangle areas and heights) and 22 indices were calculated. Two datasets were assembled including the linear measurements only and all measurements and index, respectively. An additional third dataset comprising all possible interlandmark distances between the landmarks was constructed. Two machine learning algorithms-SVM and ANN and a traditional statistical analysis LR-were applied to generate models for sex estimation. In addition, two advanced attribute selection techniques (Weka BestFirst and Weka GeneticSearch) were used. The classification accuracy of the models was evaluated by means of 10 × 10-fold cross-validation procedure. All three methods achieved accuracy results higher than 95%. The best accuracy (96.1 ± 0.5%) was obtained by SVM and it was statistically significantly higher than the best results achieved by ANN and LR. SVM and ANN reached higher accuracy by training on the full datasets than the selection datasets, except for the sample described by the interlandmark distances, where the reduction of attributes by the GeneticSearch algorithm improved the accuracy.

Sex assessment with the radius in Portuguese skeletal populations (late 19th - early to mid 20th centuries).

The assessment of sex is of immense relevance in the analysis of human skeletal remains, as other parameters of the biological profile are usually sex-specific (e.g., age at death or stature). The radius can be used to estimate sex when more dimorphic bones are not available or in the case of incomplete and fragmentary remains. Ten radius measurements collected in a sample of 364 individuals (166 females and 198 males) from the Coimbra Identified Skeletal Collection (late 19th - early 20th centuries) were employed to generate univariable and multivariable models for sex estimation. All models were evaluated with a 10-fold cross-validation method and an independent holdout sample from the Luís Lopes Collection (late 19th - mid 20th centuries) encompassing 50 individuals (25 females and 25 males). Univariable models show an accuracy ranging from 77.7% to 89.8% (cross-validation), and from 70% to 86% (test sample), while accuracy in the multivariable models varies from 88.7% to 93.4% (cross-validation), and 84.0% to 90.0% (test sample). Results suggest that measurements of the radius are useful to develop standard guidelines for sex estimation of anonymous skeletal remains.

Data mining for sex estimation based on cranial measurements

The aim of the present study is to develop effective and understandable classification models for sex estimation and to identify the most dimorphic linear measurements in adult crania by means of data mining techniques. Furthermore, machine learning models and models developed through logistic regression analysis are compared in terms of performance. Computed tomography scans of 393 adult individuals were used in the study. A landmark-based approach was applied to collect the metric data. The three-dimensional coordinates of 47 landmarks were acquired and used for calculation of linear measurements. Two datasets of cranial measurements were assembled, including 37standard measurements and 1081 interlandmark distances, respectively. Three data mining algorithms were applied: the rule induction algorithms JRIP and Ridor, and the decision tree algorithm J48. Two advanced attribute selection methods (Weka BestFirst and Weka GeneticSearch) were also used. The best accuracy result (91.9 %) was achieved by a set of rules learnt by the JRIP algorithm from the dataset constructed by application of the GeneticSearch selection algorithm to the dataset of standard cranial measurements. The set consisted of five rules including seven cranial measurements. Its accuracy was even better than the classification rates achieved by the logistic regression models. Concerning the second dataset of nonstandard measurements, the best accuracy (88.3 %) was obtained by using classification models learnt by two algorithms – JRIP with a dataset preprocessed by the BestFirst selection algorithm and Ridor with preprocessing by the GeneticSearch selection algorithm. Our experiments show that for the two datasets mentioned above the rule-based models contain smaller sets of rules with shorter lists of measurements and achieve better classification accuracy results in comparison with decision tree-based models.

Tailored logistic regression models for sex estimation of unknown individuals using the published population data of the humeral epiphyses.

This paper introduces a far-flung approach to formulate population independent models based on the humeral epiphyses as a supplementary tool for biological sex estimation of unknown partial remains. Resources for this study include the published summary statistics of 7 modern populations inhabited the continents of Africa, Asia, Europe, and South America. The regenerated humeral metric data (n=1490) via truncation approach were modeled using logistic regression. Three fitted models were evaluated for applicability across populations on an independent test sample (n=430). The experiment was assessed graphically and quantitatively using histogram of posterior probabilities and the classification table. The predictive power of the models was evaluated at the conventional (0.5) and high (0.95) posterior probability thresholds. It was found that the vertical humeral head model is insufficient for sex estimation especially in the European females due to different levels of interpopulation size variability. Interestingly, the distal biepicondylar breadth model showed overall better performance achieving the highest total and sex specific accuracies. Findings indicated that together, the epiphyseal measurements are capable of discriminating sex with overall accuracy of 90.2% which is raised up to 98.8% with 95% confidence of accurate estimates in more than 50% of the test sample. While evidences have been presented pointing to the biological and statistical meaningfulness of the humeral epiphyses model, the analysis allowed pinpointing the utility of the distal biepicondylar breadth model in sex diagnosis in transpopulation application settings. Additionally, few variables are needed to reach satisfactory sex prediction in a diverse sample.

Adjusting posterior probabilities to meet predefined accuracy criteria: A proposal for a novel approach to osteometric sex estimation

The osteometric methods are the most reliable way to estimate the sex of skeletons when DNA analysis is not used. However, as osteometric studies usually ignore the overlap in male and female skeletal dimensions, they rarely achieve accuracy sufficient for forensic application. To resolve this issue, recent studies suggest sex estimation only when posterior probability (pp) is greater than 0.95, but that approach does not always provide sufficient accuracy and creates a large proportion of unsexed skeleton. Thus, our study aimed to explore whether it is possible to adjust pp on skeletal measurements with pronounced sexual dimorphism to meet 95% accuracy and to enable sex estimation on a reasonable proportion of individuals. From 207 skeletons, we included 65 postcranial measurements and selected 10% of variables with the highest sexual dimorphism. We computed univariate and bivariate discriminant functions using pp threshold of 0.5, 0.95, and the threshold required to achieve accuracy of ≥ 95%. Discriminant functions with pp=0.5 obtained accuracy of 85%–93%, while those with pp≥0.95 and adjusted posterior probabilities obtained 94%–99%. However, we showed that by selecting a particular threshold, sex could be estimated on a greater proportion of individuals than for pp≥0.95: 42%–86% vs. 24%–62% for univariate and 69%–95% vs. 49%–78% for bivariate functions. Therefore, when developing sex estimation models, we suggest not to use fixed pp level, but to adjust pp to achieve 95% accuracy and to minimize the percentage of unsexed skeletons.