Curve Fitting
1. For linear curve the given data has a lot of deviation from the given values, thus it is not a good solution as the predictability of this curve have large deviation from the actual answer.
For cubic curve the given curve is almost satisfied thus is a good solution as it has a better predicatability.
2. The curve fit can be made perfect by either increasing the degree of polynomial of the curve or using different polynomial for each domain
![\"curve](\"//edxengine-live-courses-files.s3.amazonaws.com/1534872385.jpg\")
%inputs parameters
cp_data = load(\'data\');
temparature = cp_data(:,1);
cp = cp_data(:,2);
% curve fitting
% a*t^2+b*t+c
co_effs = polyfit(temparature,cp,2);
predected_cp = polyval(co_effs,temparature);
% comparing original data and curve fitting
hold on
plot(temparature,cp,\'linewidth\',3);
plot(temparature,predected_cp,\'linewidth\',3,\'color\',\'red\');
xlabel(\'tempareture [k]\');
ylabel(\'specific Heat [KJ/K.mole-K\');
legend(\'original dataset\',\'curvefit\')