Context-based Marwari text clustering

Today's news is not only produced quickly and in large quantities, but it also originates from a variety of internet sources, including news organizations and websites that help people find jobs. It is usual practice to extract information from English text, however, Marwari, a regional language, has not been fully investigated because of the scarcity of resources. Marwari Vector Space Model for (MVSM), Marwari Synset Vector Space Model for (MSVSM), & the Grouped Entity Marwari Document Matrix (GEMDM) were used to process over 1223 Marwari news items using entity extraction and clustering algorithms. To improve accuracy, we used synset-based dimension reduction techniques. MGEDM performed the best across several datasets when the Hierarchical Agglomerative Clustering (HAC) and K-means clustering was evaluated using three measures. MGEDM obtained a purity score of 1.8, an entropy score of 1.4, and positive silhouette’s coefficient.