Application of data mining techniques for economic evaluation of air pollution impact and control
Lukman, Iing (2007) Application of data mining techniques for economic evaluation of air pollution impact and control. PhD thesis, Universiti Putra Malaysia.
In this research we examine aspects of the interdependence between economic development and the use of environmental and natural resources assets from global data published by United Nations. For that purpose, we use data mining techniques. Data mining techniques applied in this thesis were: 1) Group method of data handling (GMDH), originally from engineering, introducing principles of evolution - inheritance, mutation and selection - for generating a network structure systematically to develop the automatic model, synthesis, and its validation; 2) The weighted least square (WLS) and step wise regression were also applied for some cases; 3) The classification-based association rules were applied. Data sets for this research consist of two sets integration data of air quality data and macroeconomic data of the cross-country data of World Development Indicator 2003 (WDI 2003), and from www.nationmaster.com. The results from www.nationmaster.com were as follows: the corruption index was strongly related to the urban SO2 concentration. The corruption index along with NOx emission has big contribution to the debt. Debt is the debt of the home country to the foreign country or external debt or foreign debt. The result from WDI 2003 shows that the mortality rate of children under five years old depended on sanitation and water facilities obtained from GMDH results. However, the results from stepwise regression shows that mortality rate was dependent on annual deforestation, particulate matter, nationally protected area, but the big contribution was from annual deforestation. Based on GMDH, new Gross National Income (GNI) formula was found. Previously GNI was known as Gross National Product (GNP). It was different from the common formula of GNP. The formula or equation model of urban SO2 concentration was also found through the GMDH algorithms. The results were then compared to WLS and Stepwise regression. The debt was found by GMDH to be dependent on the corruption index as well as urban SO2 concentration. Corruption index along with NOx emission were related to debt. Results from weighted least square using SAS software showed that the corruption index was significant to the concentration of urban SO2. Results from classification rules of the WDI 2003 data showed that the more energy imports net from foreign country was associated with the smaller in adjusted net saving in home country. Energy imports net were calculated as energy use in oil equivalents. This indicated that if the energy imports net was higher, then the adjusted net saving was small, and then CO2 emissions was small also. Thus, to reduce global warming in home country, a country can import energy from foreign country. According to the result from association rules on nationamaster.com data there were indication that corruption index was related with higher urban SO2 concentration, and inflation. Results from association rules of item sets shows that the urban SO2 always follows the direction of corruption index. In addition, if any country wants to reduce the urban SO2 concentration, more works can be conducted on controlling corruption index than controlling SO2 emission per populated area.
Repository Staff Only: Edit item detail