Article Preview
Top1. Introduction
In recent years, large number of data is collected from social networks and mobile devices. Here, the data is increasing at the annual growth of 42% through 2020 according to the IDC research (Callan, 1994; Nguyen, et al., 2019). The increasing volume of the documents influences the improved interest for recovering massive data. Vast documents with miscellaneous structures and small documents with multiple subjects are the main dare in document retrieval. The MapReduce framework (Kamal, et al., 2016; Kamal, et al., 2017) is a parallel programming approach that may be introduced based on the Hadoop environment. It comprises of Mapper and Reducer phase that are customized based on the user’s need for data processing. In this case, the input data are partitioned into chunks of data, and then these chunks are forwarded to worker node for processing (Thuy, et al., 2019; Rajendran, et al., 2018).
After reaching the worker nodes, the Mapper function processes the (Key, Value) pairs in the input file and produces the intermediate result. The intermediate result is given as the input to reducer function that reduces the data (Rajendran, et al., 2018). Clustering is the data analysis tool to identify the structure of pattern and the information of unlabelled data. It is a data analyzing scheme that systematizes the group of patterns as clusters based on their similarity (Hsu, 2006; Hsu & Huang, 2008). Data mining is the method for extracting valuable veiled knowledge from the enormous data sets (Can, et al., 1995; Chakraborty & Nagwani, 2011). Nowadays, incremental clustering becomes very popular for meeting the demand for online applications.
The document clustering is very helpful for both searching and browsing, while this may be luxurious for huge collections (Thuy, et al., 2019). The Incremental clustering approaches work by processing the data once at the time, and assigning the incremental data instances to their related clusters when they progress. Some of the clustering methods including hierarchical-based incremental clustering (Widyantoro, et al., 2002; Zhao, et al., 2018), incremental affinity propagation (IAP) clustering (Sun & Guo, 2014; Zhao, et al., 2018), K-centroids-enabled incremental clustering (Chakraborty & Nagwani, 2014; Zhao, et al., 2018), density-driven incremental clustering (Ester, et al., 1998 ; Zhao, et al., 2018), soft and fuzzy-based incremental clustering, and Clustering by fast search (CFS) (Zhao, et al., 2018). The possible application areas of clustering include: information retrieval (Diamantini, et al., 2013), image analysis (Thomas & Rangachar, 2018), and in the domains, such as business, economics, chemistry (Cannuccia, et al., 2011) and so on.