The line in the option C is the best model for the data in the scatter plot. Since in this scatter plot the points above and below the line are equally distributed (i.e. there are 5 points above the line and 5 points below the line) whereas in the rest three graphs the distortion above and below the graph is unequal.
