MULTIPLE REGRESSION ANALYSIS



The dependent variable bu68 (built up area for the year 1968)
Independent variables
Pop 68 (population of each grid)
Dispt (the distance from Municipalities)
Disrd (the distance from highway)
Discbd (the distance from CBD)
Unlike 1968, there can be a change of trend as seen in the built up 2002 map. Urban area based on municipalities started sprawling away from the municipalities and the relevance of Central Business District has started gaining momentum.
Thus for 2002 scenario
The dependent variable bu02(built up area for the year 2002)
Independent variables
Pop 02 (population of each grid)
Discbd (the distance from CBD)
Dispt (the distance from Municipalities)
Disrd (the distance from highway)
1968 model
bu68=-34.6731+6.981579lnpop68-0.03441discbd-0.37763dispts-0.54502disrd
2002 Model
bu02=-78.4362+14.70746lnpop02-0.32722discbd-0.28643dispts-1.06272disrd
Comparison of the regression results 1968 and 2002
1. Two variable regression analysis gives significant P values for all variables.
2. However, inclusion of all independent variables gives p value for two independent variable insignificant results. This reflects the multicolinearity problem.
3. However combining all these gives a better R2 value and hence all independent variables are included in the final model for 1968 and 2002
Between the 1968 and 2002 ,the significance of dispt is lost and discbd has gained. This is an indication that the development of Kochi Metropolitan Area has its focus changed from municipality concentrated to CBD concentrated. Greater Kochi Metropolitan Area is emerging as a big city focused on CBD and hence intervention is required to attain a polynodal urban structure on sustainability considerations.