Big data stream classification in Apache Spark platform using adaptive dragonfly moth search algorithm

Big data streaming is done using two phases, offline and online, which is carried out based on the proposed optimisation algorithm, named adaptive-DMS algorithm. In the offline phase, the input text data is initially classified as sub-sets and provided as the input to individual slave nodes. In the slave nodes, the pre-processing is done to remove the unwanted data's present in the input using stop word removal and stemming. After pre-processing, TF-IDF is applied for extracting the best features and then classification is done. The same process is repeated for online phase. The error is determined based on the resulted features obtained from online phase and offline phase. If the error is maximal, the final classified data is determined by remodel the classifier by setting the boundary weights.