FlexPro Inductive Statistics Option
Introduction
The Inductive Statistics Option offers you a variety of statistical tests and the possibility of calculating theoretical distributions.
Unlike with descriptive statistics, which permits a description of the
data material, using statistical characteristic quantities, the
inductive statistics testing and estimating procedures offer the
possibility, based on samples, of classifying the population from which
these samples originate. An important domain of use for inductive
statistics is the statistical process control (SPC).
Features
- Goodness-Of-Fit Tests: Chi-Square test with adjustable number of classes and Kolmogoroff-Smirnov test for normal and exponential distribution. The error probability can be specified for both tests. The parameters for the distributions can be estimated or specified.
- ANOVA: Square Sum Of Treatments (SST), Mean Square Sum Of Treatments (MST), Square Sum Of Errors (SSE) , Mean Square Sum Of Errors (SSE), Total Square Sum (SSG). The error probability can be specified.
- Outlier Correction and Outlier Tests: David-Hartley-Pearson test und Grubbs-Beck test with adjustable error probability.
- Variance Tests: Bartlett test and F test with adjustable error probability.
- Distributions: Continuous distributions normal, log-normal, exponential and Weibull. Test distributions chi-square, student-t and F. Discrete distributions binomial and Poisson. The parameters for the normal and exponential distributions can be estimated from a sample. With continuous and test distributions, the density and the distribution function can be computed, both normalised to one or hundred. With discrete distributions, the density function can be computed.
- Confidence Intervals: Confidence intervals for the parameters variance and mean value of the normal distribution for a given confidence level.