Jadwal Sholat

Memuat jadwal sholat…

Ilmu Sosial editorial

Open AccessOA2022

Performance Analysis of Support Vector Machine in Sex Classification of The Sacrum Bone in Forensic Anthropology

This study evaluates the performance of Support Vector Machine (SVM) with linear, polynomial, and RBF kernels in classifying the sex of the sacrum bone using six metric variables. The polynomial kernel achieved the highest accuracy of 85.56%, followed by the linear kernel at 84.44% and the RBF kernel at 83.33%.
Iis Afrianty; D. Nasien; H. Haron· JURNAL TEKNIK INFORMATIKA· 2022· DOI 10.15408/jti.v15i1.25254

The core problem

Sex classification is a fundamental task in forensic anthropology, aiming to determine whether skeletal remains belong to a male or female. The sacrum bone, due to its morphological differences between sexes, is a valuable element for identification. This paper investigates the application of Support Vector Machine (SVM) for sex classification of the sacrum using metric measurements. The study focuses on six variables: superior breadth, anterior length, mid ventral breadth, real height, diameter of the base, and max-transverse diameter of the base. The objective is to analyze the performance of SVM with different kernel functions—linear, polynomial, and radial basis function (RBF)—and to identify the kernel that yields the highest classification accuracy.

Innovation

The dataset consists of sacrum bone measurements from individuals of known sex. Six metric variables were recorded: superior breadth, anterior length, mid ventral breadth, real height, diameter of the base, and max-transverse diameter of the base. The SVM implementation used the libSVM library. Three kernel functions were evaluated: linear, polynomial, and RBF. For each kernel, a grid search was performed to optimize hyperparameters: the regularization parameter , the kernel coefficient (for polynomial and RBF), and the polynomial degree . The performance metric was classification accuracy. The experimental setup involved training and testing on the dataset, with results reported for the best hyperparameter settings per kernel.
Introduction
Sex classification is a fundamental task in forensic anthropology, aiming to determine whether skeletal remains belong to a male or female. The sacrum bone, due to its morphological differences between sexes, is a valuable element for identification. This paper investigates the application of Support Vector Machine (SVM) for sex classification of the sacrum using metric measurements. The study focuses on six variables: superior breadth, anterior length, mid ventral breadth, real height, diameter of the base, and max-transverse diameter of the base. The objective is to analyze the performance of SVM with different kernel functions—linear, polynomial, and radial basis function (RBF)—and to identify the kernel that yields the highest classification accuracy.

Why it matters

The polynomial kernel's superior performance suggests that the relationship between the sacral measurements and sex is better captured by a polynomial decision boundary. The linear kernel's performance is close, indicating that the data may be nearly linearly separable. The RBF kernel, despite its flexibility, did not achieve the highest accuracy, possibly due to overfitting or suboptimal hyperparameter tuning. The study demonstrates that SVM is a viable tool for sex classification from sacrum measurements, with accuracies exceeding 83% for all kernels. However, the relatively small dataset size may limit generalizability. Future work could include larger samples and additional variables. The results are consistent with previous studies in forensic anthropology that have used SVM for sex determination from other bones.

Who should read this

Pembaca Research Digest dan praktisi rumpun ini.

Opening member content…