MACHINE LEARNING FOR SPATIAL DATA

Nowadays, machine learning, artificial neural networks, and support vector machines are important tools for spatial and environmental data analysis and visualization. Machine learning for spatial data is used to predict and classify unknown locations according to known locations. Spatial data exist in different formats like vector data, raster data, geographic co-ordinate system. Usage of spatial data has been increased in recent times. Geographic information system (GIS) is the most common way to analyse the spatial data. Spatial data is used to track infectious diseases, climate change simulations, etc., There are many issues in spatial data handling like huge amounts of available data, missing data, duplicate data, classification of spatial images. This paper presents the review about the classification of images in a dataset using machine learning algorithm and finds the accuracy of prediction between support vector machine and Convolutional neural networks. The intersection of machine learning and spatial data involves four steps, including analysis and applicability assessment, extending the algorithm by embedding spatial properties, tuning parameters for good results and extending the algorithm in multiple ways. From this study, we learned about the spatial properties and algorithms of machine learning with spatial data. Machine learning methods are more effective for spatial image classification and have wide applications these days.