ETHNICITY CLASSIFICATION USING A DYNAMIC HORIZONTAL VOTING ENSEMBLE APPROACH BASED ON FINGERPRINT

Today, there is a fierce rivalry between ethnic groups in Nigeria on a number of issues, such as the division of power and resources, aversion to dominance, and uneven growth. Ethnicity as an identity naturally occupies a prominent position in the political arena. It is the simplest and most natural way for people to mobilize around essential human needs such as security, food, shelter, economical well-being, inequity, land distribution, autonomy, and recognition. Recent research has revealed the potential to determine an individual's ethnicity based on biometric data automatically. These studies reported significant advancements in automatically predicting demographics based on facial and iris traits. This success has been ascribed to the availability of a sufficient amount of high-quality data. There needs to be more data about the likelihood that fingerprints can disclose an individual's ethnicity. A need for more data causes this difficulty. This study aims to obtain fingerprint pictures via live scan among the major ethnic groups in Nigeria. For training and classification of the fingerprint images, the proposed Dynamic Horizontal Voting Ensemble (DHVE) deep learning with a Hybrid of Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) as the base learner was employed. Standard performance classification metrics such as Accuracy, Recall, Precision, and F1 score were used to evaluate the performance analysis of the model. This study demonstrated an accuracy of over 98% in predicting a person's ethnicity. Additionally, the proposed model outperformed existing state-of-the-art models.