A New Modified Extended Incremental Clustering Algorithm to Handle Real-Time Dynamic Data

In this research paper, we proposed a new incremental clustering algorithm to handle dynamic datasets, data samples are processed like a batch, and our method allowed us to insert or detect data samples at any point of time. Data samples are classified into different clusters based on Euclidean distance and similarity value. All data samples are classified into core, noncore, or noisy data points. In case of insertion of new data samples, the algorithm repeats the procedure to cluster data samples into different clusters. For deletion, data samples are removed from existing clusters as well as from the original dataset. For new data sample insertion checking any impact on existing clusters is more important and are there any impact recomputed core points. Input dataset size may be increased or decreased then the complete procedure needs to be repeated as a batch of data samples. Traditional clustering algorithm time complexity is O (n2) for static datasets and in our proposed method it is reduced to O (n) for dynamic datasets.