Determination of Gender from Pelvic Bones and Patella in Forensic Anthropology: A Comparison of Classification Techniques

Abstract

The determination of gender is an important part of forensic anthropology because as the first essential step for positive identification process. Besides empirical methods for gender determination such as Discriminant Function Analysis (DFA), Artificial Intelligence methods such as Artificial Neural Network (ANN) should be considered to obtain more accurate determination result. This paper proposes Back propagation Neural Network (BPNN) model of ANN methods. By using data and DFA result of pelvic bones and patella from previous work, this paper compares accuracy of result obtained from the BPNN models. A total sample data of 136 pelvic bones and 133 patellae have been collected. For pelvic bones, BPNN gave average accuracy as much as 98.5% for training and 98.3 for testing. While on left pelvic bones, average accuracy that is obtained are 98.49% for training and 86.6% for testing. For patella bones, all average accuracy (males and females) are obtained by BPNN is 94.09%. If compared with previous study that using DFA obtained accuracy as much as 92.9%. It is concluded that in gender determination, BPNN gives high accuracy of classification for both bones compared with DFA.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 18 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.