Reasoning Temporally Attributed Spatial Entity Knowledge Towards Qualitative Inference of Geographic Process

Reasoning Temporally Attributed Spatial Entity Knowledge Towards Qualitative Inference of Geographic Process

Jayanthi Ganapathy (Anna University, Chennai, India) and Uma V. (Pondicherry University, Puducherry, India)
Copyright: © 2019 |Pages: 22
DOI: 10.4018/IJIIT.2019040103

Abstract

Knowledge discovery with geo-spatial information processing is of prime importance in geomorphology. The temporal characteristics of evolving geographic features result in geo-spatial events that occur at a specific geographic location. Those events when consecutively occur result in a geo-spatial process that causes a phenomenal change over the period of time. Event and process are essential constituents in geo-spatial dynamism. The geo-spatial data acquired by remote sensing technology is the source of input for knowledge discovery of geographic features. This article performs qualitative inference of geographic process by identifying events causing geo-spatial deformation over time. The evolving geographic features and their types have association with spatial and temporal factors. Event calculus-based spatial knowledge formalism allows reasoning over intervals of time. Hence, representation of Event Attributed Spatial Entity (EASE) Knowledge is proposed. Logical event-based queries are evaluated on the formal representation of EASE Knowledge Base. Event-based queries are executed on the proposed knowledge base and when experimented on, real data sets yielded comprehensive results. Further, the significance of EASE-based spatio-temporal reasoning is proved by evaluating with respect to query processing time and accuracy. The enhancement of EASE with a direction for further development to explore its significance towards prediction is discussed towards the end.
Article Preview
Top

1. Introduction

Knowledge Engineering (KE) is the subject of Artificial Intelligence (AI) in which Knowledge Representation and Reasoning (KRR) is conceptualized by the formal representation of facts. Monitoring earth resources at temporal scale by remote sensing technology helps in acquiring data pertaining to Geo-spatial dynamism Campelo et al. (2012). Furthermore, Geo-spatial data acquired at varying time instances represent the facts that describe the spatial phenomena occurring at various geographical locations. Computations on such facts based on the principles of KE Patterson (2005) yields new knowledge about Geo-spatial changes occurring over time. Qualitative Knowledge Representation and Reasoning govern the representation of Geo-spatial dynamism. An activity in a geographic region taking place at an interval and that describes a spatial change is a Geo-spatial event. Such an event occurring consecutively in the temporal range is referred to as Geo-spatial process. Eventually, these computations are the essentials of geomorphology Jayanthi and Uma (2018). Apparently, the phenomenal change resulting due to critical deforestation activities like forest fire, logging and cattle ranching causes land cover deformation in geographic regions such as forest, farmlands, landscapes etc. Automated reasoning involves qualitative analytics on spatial relations. Temporal description of such spatial relations essentially provides comprehensive results in qualitative prediction of Geo-spatial dynamism. Thus, information processing towards qualitative analysis on such facts involves logical computation on spatial relations by considering the temporal factors (Galton et al., 2005; Kang et al., 2016; Liu et al., 2015). This work aims at investigating Spatio-Temporal Knowledge Representation and Reasoning (STKRR) approaches towards qualitative analysis of Geo-spatial dynamism. Further, a knowledge based geographic process inference system is proposed.

In our previous works, a comprehensive survey of various approaches in representation of geographic facts, considering spatial and temporal factors is presented Jayanthi and Uma (2017). Rudimentary modeling of geographic spatio-temporal facts by formulating event attributed spatial entity knowledge and hypothetical evaluations of queries were illustrated in our previous works Jayanthi and Uma (2017 and 2016). This work concentrates on proving the consistency of the well formulated knowledge base by deducing the answers to queries logically. Additionally, in this work, evaluation of the geographic process inference system has been performed by considering real world dataset. Contributions in this work are (1) Construction of Event Attributed Spatial Entity (EASE) Knowledge base (2) Qualitative inference of geographic events and processes.

This paper explores the significance of formal representation of temporal and spatial facts in EASE knowledge base and qualitative reasoning of such facts in inferring spatial process. The spatial processes inferred can be used for forecasting. To substantiate this, the authors have logically proved that prediction can be performed considering the events and processes at previous time instances. But, in this work the authors have not considered time series forecasting using machine learning algorithms since it is beyond the scope of this work. This work explores the significance of EASE in inferring geographic events and spatial process thereby, illustrates the importance of EASE in inferring spatial process using qualitative analytic method.

The details on organization of the paper are as follows: second section enumerates on KRR approaches and related works found in the literature. The third section introduces the Geographic process inference system and presents the formal representation of KBEASE. Fourth section elaborates the logical evaluation of queries. Fifth section illustrates the experimental analyses on real datasets followed by empirical evaluation of the system which is discussed in sixth section. Seventh section enumerates the significance of KBEASE and concludes the work with direction to enhance further.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 16: 4 Issues (2020): 1 Released, 3 Forthcoming
Volume 15: 4 Issues (2019)
Volume 14: 4 Issues (2018)
Volume 13: 4 Issues (2017)
Volume 12: 4 Issues (2016)
Volume 11: 4 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing