Covering Rough Clustering Approach for Unstructured Activity Analysis

2. Related Works

Data are generated and collected using sensors, cameras and smart devices. The generated data create vagueness, incompleteness, and granularity in an information system which gives unreliable solution in the data analysis. In our day to day approach, several traditional tools are used for formal modeling, computing and reasoning of data, which are basically crisp and deterministic in nature. Real situations are very often not crisp and deterministic, and they cannot be described precisely. Rough Sets is one of the excellent modeling tool introduced by Pawlak (Pawlak, Z.,1982,1991) which is used to study the imprecise data in the information system. Applying rough sets in real-time information systems leads to some restriction because of the equivalence relation unless the clustering of problem appears to hold true for equivalence relation.The equivalence relations of rough sets were extended to generalized binary relations in several directions. Similarly, partitions of the universes used to define rough sets were extended to coverings. A type of generalized rough sets based on covering and the relationship between this type of covering- based rough sets (Zhanhong Shi & Zengtai Gong, 2010) and the generalized rough sets based on binary relation were studied. Clusters can be hard or soft in nature. In soft clustering, an object may be a member of two or more clusters. Soft clusters may have fuzzy or rough boundaries. Web mining is one such area where overlapping clusters are required. Generally, clustering algorithms make use of either distance functions or similarity functions for comparing pairs of sequences.