Imputation using normal distribution sampling

[henninggaertner]

A new type of imputation should also be added. Here, the NA values are replaced by random values from a normal distribution. The normal distribution is modified by a downshift and a width by the user. Downshift is the new mean value of the distribution, which is shifted to the left by x units of the standard distribution of the entire data set [mu_impute = mu_dataset + x* sigma_dataset]. The width is the factor y by which the standard distribution of the entire data set is scaled
[sigma_impute = y * sigma_dataset].

Acceptance Criteria

• the normal distribution is calculated correctly
• the imputed values are sampled correctly from the normal distribution
• the imputation method is documented
• the imputation method is available in the UI
• tests have been created