Spatiotemporal knowledge mining Framework for computer assisted Decision making system

Abstract

Advances in computer vision, social media and location-sensing technologies have accelerated the emergence of spatiotemporal data, emphasizing the importance of developing effective methods for the analysis of spatiotemporal patterns. Spatiotemporal data represents data that is collected considering spatial (space) and temporal (time) aspects of each observation. Spatiotemporal event data and spatiotemporal moving-object trajectory data are generally two types of spatiotemporal data. (George et al. 2021; Junming et al. 2014). Spatiotemporal event data represents the temporal observation of geographical or spatial components. Geo-tagged tweets, YouTube comments and movie reviews are a few examples of spatiotemporal event data. Spatiotemporal moving-object trajectory data relate to the spatial properties of objects that typically change their location, direction, size and shape over time. Location and direction are used for the capture of movement objects that can be modelled as geometric transformations such as translation and rotation. Vision-based moving data such as human-robot interaction and smart video surveillance are a few examples of spatiotemporal moving-object trajectory data. Spatiotemporal data contains hidden knowledge which can be used in building a computer assisted decision making system. newlineThe process of discovering interesting patterns and knowledge mining from spatiotemporal data is known as spatiotemporal knowledge mining. This research work proposes an effective mining framework that aims at building classification models for spatiotemporal event data and spatiotemporal moving-object trajectory data. These classification models are used in developing Computer Assisted Decision Making System (CADMS) to assist decision-making activities. Several research studies taken up so far relate to the investigation of the significance of spatiotemporal knowledge mining newline