Spatial data mining on literacy rates and educational establishments in Bangladesh

Data mining is the process of extracting non-trivial patterns from large volume of data. It generates insight and turns the data into valuable information. A critical yet common flaw when performing data mining is to ignore the geographic locations from where the data is taken. When this geospatial attribute of the data is taken into consideration, the process is known to be geospatial data mining. This task essentially deals with the detection of spatial patterns in the data, the formulation of hypotheses and the assessment of descriptive or predictive spatial models. Spatial data mining could provide interesting and useful information to government, environmentalists and relevant decision makers´ in the assessment of the relative performance of a particular geographic area. The results could also be used for causal analysis by domain experts. In our research we perform spatial data mining using literacy rates and the number of educational establishments. The data is from the 64 well defined administrative units of Bangladesh known as Zilas. This paper contains a summary of the theory, methodology and detailed analysis of results. We compare the results found by spatial model with classical regression model. The results demonstrate that spatial lag model outperforms the classical model in different perspectives.