Efficient Education Community Detection Using Deep Learning Algorithm for Similarity Identification in Online Social Networks

Education development is tremendous in online social network through information communication technology by group of various communities depends on the education types. Increasing online learners through distance education carry the search groups to find the suitable courses in the online community forms. So, community detection is important aspect for identifying the resource dependencies properties to make community groups on relevant education groups. The network of communities in the online are so huge and the information sharing among the field of entities in the groups contains various features. By identifying the communication relation feature e properties is big problem in community detection. Most of the prior Machine Learning (ML) models failed to analysis the feature relation properties lead poor accuracy because of non-relation community feature margins are grouped. To addressing the problem, to implement a Hyper spectral feature selection with deep Recurrent convolution Neural network (DRCNN). First, to normalize academic community data records collected from online community forums, Box-cox normalization is preprocessed. The community pattern interest rate (CPIR) is estimated based on the learner interest specific terms of subjectivity relation. By selecting the properties of the feature in hyper level by creating patterns and entity relation using spider ant colony Correlation Exhaustive Feature Selection (CEFS) to identify the feature relation related specific community group. Then deep Recurrent convolution Neural network (DRCNN) is attained to identify the community groups effectively. Then proposed system produce high performance compared to the system as well to find the communities properties effectively in high true positive rates to improve the performance with redundant time complexity.