Related Experiment Video
Updated: Dec 20, 2025

A Psychophysics Paradigm for the Collection and Analysis of Similarity Judgments
Published on: March 1, 2022
Multi-element comparisons of tapes evidence using dimensionality reduction for calculating likelihood ratios
Anjali Gupta1, Claudia Martinez-Lopez2, James M Curran1
1Department of Statistics, University of Auckland, Auckland, New Zealand.
This study introduces a new method using principal component analysis (PCA) to simplify multi-element forensic evidence analysis. The approach effectively calibrates likelihood ratios (LRs), improving accuracy and reducing errors in comparing samples like electrical tapes.
Area of Science:
- Forensic Science
- Analytical Chemistry
- Statistics
Background:
- Computing likelihood ratios (LRs) for multi-element forensic evidence is challenging due to data complexity.
- Existing multivariate random effects models can be unstable and produce extreme LR values, especially with more variables.
- High-dimensional data in forensic analysis requires robust dimensionality reduction techniques.
Purpose of the Study:
- To evaluate a novel method combining principal component analysis (PCA) and post-hoc calibration for analyzing high-dimensional, multi-element forensic evidence.
- To assess the performance of this method in calculating and calibrating likelihood ratios (LRs) for electrical tape samples.
- To determine the optimal number of principal components (PCs) for accurate and reliable forensic comparisons.
Main Methods:
- Applied additive log-ratio transformation to 18-element electrical tape data (analyzed by LA-ICP-MS).
- Utilized scores of the first five principal components (PCs) as input for the likelihood ratio formula (LR MVN).
- Implemented a post-hoc calibration method to adjust calculated LRs for improved accuracy and range.
Main Results:
- Initial LR calculations yielded extreme values, necessitating post-hoc calibration.
- The calibrated LRs fell within an appropriate range, demonstrating the effectiveness of the calibration step.
- The scenario using the first five PCs (LR5) achieved the lowest false exclusion (2.2%) and false inclusion (3.7%) error rates, indicating optimal performance.
Conclusions:
- The proposed method effectively reduces dimensionality in multi-element forensic data, overcoming limitations of traditional approaches.
- PCA combined with post-hoc calibration provides a stable and reliable way to compute well-calibrated likelihood ratios.
- This approach enhances the accuracy of forensic comparisons, particularly for complex datasets like those from electrical tapes.
More Related Videos
09:49Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm
Published on: December 24, 2015
08:45Mapping Cortical Dynamics Using Simultaneous MEG/EEG and Anatomically-constrained Minimum-norm Estimates: an Auditory Attention Example
Published on: October 24, 2012
Related Concept Videos
Multiple Comparison Tests
It would be easy to compare two samples using a significance alpha level of 0.05. In other words, there is only one sample pair to be compared. However, it would be difficult to identify a significantly different sample if the number...
Bonferroni Test
The means of different samples are first paired in all possible combinations.
The null hypothesis of the...
Dimensional Analysis
Conversion Factors and Dimensional Analysis
The unit...
Dimensional Analysis
Dimensional Analysis
Dimensional analysis allows us to analyze and compare physical quantities on a...
Dimensional Analysis
In fluid mechanics, dimensional...