Neighborhood Rough-Sets-Based Spatial Data Analytics

Neighborhood Rough-Sets-Based Spatial Data Analytics

Sharmila Banu K. (VIT University, India) and B. K. Tripathy (VIT University, India)
Copyright: © 2018 |Pages: 10
DOI: 10.4018/978-1-5225-2255-3.ch160


Rough Set Theory partitions a universe using single layered granulation. The equivalence classes induced by rough sets are based on discretised values. Considering the fact that the spatial data are continuous at large, discretising them may cause loss of data. Neighborhood approximations can lead to closely related coverings using continuous values. Besides, the spatial attributes also need to be given due consideration and should be handled unlike non-spatial attributes in the process of dimensionality reduction. This chapter analyses the use of Neighborhood rough sets for continuous data and handling spatially correlated attributes using rough sets.
Chapter Preview


Mining spatial data is useful in fields like weather forecasting, natural calamity prediction, crime management, transmission and spread of infectious diseases and others. This calls for expertise in these areas and the nature of spatial data. For example, representing topology in spatial data modelling is inherent to dealing with uncertainties. And, Rough Sets have been used to deal with uncertainty in spatial data mining. Pawlak’s (1982) Rough Set Theory (RST) has been used to model spatial regions with unclear boundaries. Beaubouef and Petry (1994) have demonstrated the use of rough sets have been used to query crisp data in relational databases. The Region Connection Calculus (RCC) proposed by Randell & Cohn (1992) and Egg-Yolk models by Cohn and Gotts (1996) have been blended with the approximation concepts of RST to identify vague region boundaries. Rough sets have been used by Bai et. al (2010) to identify villages with birth defects, Ahlqvist (2005) for spatial classification and analysis, Leung et. al. (2007) for discovering classification rules in remote sensor data, Øhrn A (1999) for disease diagnosis and outcome prediction and Thangavel and Pethalakshmi (2006) for dimensionality reduction.

Figure 1.

A sample geographic region

Table 1.
Sample Attribute Data of the region in Figure 1

Key Terms in this Chapter

Rough Sets: An extension of classical set theory, which provides an approximation of crisp set in terms of lower and upper and approximation.

Spatial Auto Correlation: The values of a spatial attribute tend to be close to each other and vary gradually from core to periphery of a geographic region.

Neighborhood Systems: They provide granulation structure for each element of a universe.

Distance Measure: Measures to calculate distance between numerical, categorical or mixed data.

Granulation: It is using groups or clusters of data objects formed on the basis of similarity in an incomplete information system.

Complete Chapter List

Search this Book: