The Use of Principal Component Analysis to Provide Objective Methods for the Examination of Arabic Signatures

The natural variation within an individual’s Arabic signature is studied using Principal Component Analysis (PCA) with orthogonal rotation of the components. Fifty replicate analyses of sixteen features, measurable or classified by a numerical coding system, were analysed for a legible and an illegible type of signature. With either type of signature this PCA method provides an objective, robust and simple routine for identifying the features and importantly associations between features. Using only the measurable features, the results from two trials proves the method could potentially assist in confirming the authenticity and identifying forgeries for both legible and illegible types of signatures. From studying the effect of sample size it is recommended that no less than twenty-five control signatures should be used to ensure reproducibility and confidence in reporting any results when using this statistical method of analysis.