Gaussian KDE with Scott's rule, Silverman's rule, and fixed bandwidth —
mirrors scipy.stats.gaussian_kde with pdf, evaluate,
integrate, cdf, and resample.
Create a KDE from data and evaluate it on a grid. The factor property exposes the bandwidth.
Choose between "scott" (default), "silverman", or a fixed numeric bandwidth. Use resample to draw new samples.
integrate(a, b) gives probability mass in an interval. cdf(x) gives the cumulative distribution up to x.