Explorations on Influential node Identification and Gender Prediction in Dolphin Network through Social Network Analysis

Exploration of latent information in Social Networks through computational methods has spanned enormous heights in the recent years revealing interesting facts on diverse networks involving both animate and inanimate beings. Dolphin Network data is one such Social Network that entails new-fangled data. This research work is focused on detecting new patterns pertaining to disclosure of influential dolphins, Gender Prediction and Outlier Identification in the Bottlenose Dolphin network using Social Network Analysis. This paper proposes an alternative way to identify influential nodes with high centrality by coalescing the perceptions used in betweenness centrality measure and Random Walk measure. Gender Prediction is done through interaction based mining. Association based outlier analysis method was utilized to isolate the outliers in the data. Experimental evaluations on various social networks datasets (weighted and non-weighted) provide strong evidences for the newly proposed measures‘ utility. The proposed centrality measure is also validated for its significant difference from the existing traditional measures using one way ANOVA test followed by TUKEY HSD test and Duncan test with SPSS software package.