ENH: Add confidence interval to plot_target_feature_distribution

[ibenfjord]

ENH: Add a Bootstrap function #672 should be done before this one

Is your feature request related to a problem? Please describe.
It would give more information if the confidence intervals are visualised in the figure

Describe the solution you'd like
Use a bootstrap function to calculate the confidence intervals then use
percentile = np.percentile(final_bootstrap, (2.5, 97.5))

The percentile needs to be transformed since the bar plot function subtracts the numbers.
percentile[0, :] = abs(data - percentile[0, :]) percentile[1, :] = abs(percentile[1, :] - data)

Then add the following to the _plot_barh
xerr=percentile if n_boot else None