Text documents are a collection of informative and unstructured text characteristics that uses several clustering algorithm, which leads to misleading information with the presence of informative features. The presence of these features tends to reduce the performance of clustering algorithms to retrieve the documents within the clusters. On other hand, the study tends to get affected with high-dimensional dataset features that offers major weakness associated with the text documents. This nominally increases the execution time and reduces the overall algorithm performance. To resolve this, we present a text document clustering using multi-objective particle swarm optimization algorithm (PSO) for large datasets. The PSO is designed in order to reduce the improve performance in poor document clusters. It is designed to avoid high-dimensional dataset problem. The performance of the proposed method is validated than other existing methods against various datasets that includes: accuracy, precision, recall, normalised mutual information (NMI) and adjusted rand index. The experimental result shows that the proposed multi-objective text document PSO clustering techniques offers improved performance than ACO and QPSO algorithms.
Volume 12 | Issue 6
Pages: 2262-2269
DOI: 10.5373/JARDCS/V12I6/S20201185