Twitter Profile Information based on Feature Levels for the Identification of Individuals

Abstract

The objective of a biometric method is to make a human-like decision about identifying an individual based on physiological or potentially social characteristics. Due to the low quality of the data, biometric-based recognition schemes can be exceptionally confounded. In order for a computerized biometric framework to make decisions consistently and dynamically, Social Behavioral (SB) prompts are essential. A system that identifies an individual person on the social network (SN) is proposed in order to improve the efficiency of the traditional biometric frameworks. The social-behavioural cues of individual people were extracted from an online tweeter network. A study shows that social biometric features are more effective than biometric features based on rank. A person's score is determined by the Product Moment Correlation Coefficient (PMCC) based on the SB attributes. Furthermore, the correlation coefficient is used to calculate four different SBB features. The impact of absolute scores with their SBB model is tested directly from the OSN datasets for various kinds of subjects. Through the use of specific datasets and by comparing registered features with different types of test features, the cross-validation functionality of the system is demonstrated. The proposed system performs better than the conventional biometric framework for identifying individuals. The experimental results using SB fusion techniques demonstrated a recognition rate of 95% and were faster than most other biometric systems. Based on the SB information available on the network, the feature-based fusion technique is the best compromise, and it is commonly applied in multimodal biometric frameworks to improve precision.