An Effective Approach for Mining Complex Spatial Dataset
Grace L. Samson · University of Huddersfield Repository (University of Huddersfield) · 2012
In this research work, we have presented an illustration of spatial ecological predictive modelling.We focused on the unique features that distinguish spatial data mining from classical data mining, and presented major accomplishments of spatial data mining research, especially regarding predictive modelling, spatial outlier detection, spatial co-location rule mining, and spatial clustering.In a very detailed research based study, we thoroughly investigated methods of mining patterns of a spatial data set (which generally describes any kind of data where the location in space of object holds importance) and made predictions based on the outcome of our analyses.We based this research on the analysis of some spatial characteristic of certain objects (that exist in an ecosystem).We began with describing the spatial pattern of events or objects with respect to their attributes, in other words and most specifically, we looked at how to describe the spatial nature/characteristics of entities in an ecological environment with respect to their spatial and non-spatial attributes.Secondly, we were able to predict likelihood of an object with a range of variables (using spatial analyst tools likedistance, interpolation, overlay, raster creation, reclass, multivariate analysis, maths, surface and conditional tools respectively) on a sample dataset and then we verified the model performance on the rest of the data.These feats were basically achieved using data visualization-which is the visual interpretation of complex relationships in multidimensional dataand statistical interpretations.Results: At the conclusive end of this project, we were able to build a prediction/suitability model for the prediction of plant species around a river area.The method illustrated in this research work suggests the use of mapping and statistical functions in the prediction of large spatial database.We have tried to achieve this by two (2) major stagesfirst stage is the spatial analysis, while the second stage is the statistical analysis.The advantages and shortcomings of this approach are discussed in the context of the need for further development of methodology and software This work is particularly useful to researchers in the field of data mining as it contributes a whole lot of knowledge to different application areas of data mining especially spatial data mining.It can also be useful in teaching and likewise for other study purposes DedicationThis project work is dedicated to the almighty GOD who makes all things possible for us in his own time and my family especially my little wonderful kids, who has been a backbone to my success.