Fit data with a smooth spline; normalize to make AUC equals to 1; then calculate Jaccard distance between two curves.

Common distance matrices in optimization

Gaussian probabilistic density function is log quadratic. Laplace probabilistic density function is log linear.

Despite differences in original position, size, and orientation, Data A and B should overlap very well if they have the same "shape".

To fit an exponential (decay) model to data which has a trend of decaying and stabilizing over time.

Peak (local maxima) detection demo

To find (almost) the best span for LOESS by cross validation

Scatter plot of an absolute linear model with Gaussian noise.