Methods of Estimating Parameters of Skewed or Truncated Normal Distribution in the Presence of Observations Outside of Measurable Range
| Authors | |
|---|---|
| Year of publication | 2023 |
| Type | Conference abstract |
| MU Faculty or unit | |
| Citation | |
| Description | Every laboratory equipment has limits to what it can accurately measure. Generally, for every laboratory apparatus three types of limits should be defined – limit of blank (LoB), limit of detection (LoD) and limit of quantitation (LoQ). If an observation falls outside of the measurable range, there is an issue of estimating parameters of the distribution. In this contribution we look at four different methods applied to samples generated from skewed and truncated normal distributions – ignoring censored observations, replacing censored observations, using a truncated version of target distribution, and using target distribution with censored observations. To compare these methods we designed a simulation study, where generated samples were truncated from the left at selected quantiles. Parameters’ estimates were then compared to the original values. Simulation study was run separately on skewed normal distribution and truncated normal distribution. Based on the results we created recommendations for practical data analysis. |
| Related projects: |