Community detection is important area of research in social and educational networks. Community detection methods are usually NP-hard so one applies heuristics or approximation algorithms that have their own methodical preferences. The deficiency of trustworthy “ground-truth” makes the assessment of identified communities enormously difficult. Present community detection methods have several drawbacks. They are either based on heuristics, do not scale to large networks or do not consider overlapping community structure that is common in networks.

We propose a simple but fundamentally different view of network community detection. Networks with explicit “ground-truth” communities enable us to explore several important open questions in network community detection. We explore how ground-truth communities overlap, propose new evaluation metrics, and design novel community detection methods that more accurately detect communities and also scale to large networks.

In this work we present modified Block Two-Level Erdos-Renyi (BTER) and Chung and Lu (CL) model using weighted degree and weighted Laplacian matrices. Based on these we defined modified local, global and edge clustering coefficients. This work presents the importance of modified node degree, degree distribution, local, global and edge clustering coefficient correlation for efficient community detection (Efficient community detection and situational awareness algorithm). We test and experiment with real world, big data and large-scale technological, social and educational networks. We observe that modified node degree and clustering coefficients play an important role based on the different types of network, such as technological, social, and educational networks. Community identification is helpful in understanding the main or important components constituting a realistic (real-life) system, how these components cooperate and, finally, how they grow and influence or govern the overall network and its functions.