Page 72 - Veritas Vol 3, Issue 2
P. 72
62
It involves methods such as k-Nearest Neighbour (kNN) and Linear
Discriminant Analysis (LDA), which works on the principle of classifying
samples based on their characteristics. Support Vector Machines (SVM),
Decision Tree and Random Forests method which works on rule-based
classification, and Artificial Neural Network (ANN). Together all these
methods play their roles in applications ranging from entomology to velocity
detection and reconstruction.
Image 2: Diagram illustrating Multivariate Analysis
2.1. Regression Technique
Multivariate regression techniques being the most dependent method, it works
by establishing connections between various factors and components along with
a dependent outcome. MLR (Multiple linear Regression) one of the most
famous methods under Multivariate regression method, helps the forensic
scientist by generating linear equations of the provided datasets by
incorporating various predictor variables. PCR (Principal Component
Regression) replaces MLR as its effectiveness is limited by various factors such
VERITAS VOLUME: 3, ISSUE: 2 WRITER: R V ISHITHA REDDY
