Ensemble Assisted Multi-Feature Learnt Social Media Link Prediction Model Using Machine Learning Techniques

In this paper a robust consensus-based ensemble assisted multi-feature learnt social media link prediction model is developed. Unlike classical methods, multi-level enhancement paradigm was considered where at first the focus made on extracting maximum possible features depicting inter-node relationship for high accuracy of prediction. Considering robustness different feature sets, we extracted local, behavioural as well topological including Jaccard coefficient, cosine similarity, number followers, intermediate ADAR. The use these all link-signifier strengthened proposed link-prediction to train over large data and ensure higher accuracy. Undeniably, aforesaid multiple features-based approach could yield reliability; however, cost increased computation. To avoid it, selection methods like rank sum test, cross-correlation, principal component analysis were applied. had dual intends; assess which type can have second reduce unwanted This research revealed that similarity-based don’t significant impact eventual classification. On contrary, cross-correlation PCA based exhibited relatively (up 97%). Once retrieving set suitable features, unlike standalone classifier (two-class) prediction, designed novel consensus learning by using logistic regression, decision tree algorithm, deep-neuro computing algorithms (ANN-GD ANN-LM with hidden layers), classified each node-pair Linked or Not-Linked. Our has (98%), precision (0.93), recall (0.99), F-Measure (0.97), than other state-of-art machine methods.