Quasi-difference
In some specialist fields, the quasi-difference is the difference between two measured values or estimated values under the distorting influence of systematic errors ( trend ) or correlated measurement deviations. They can arise from the external circumstances of the measurement or from the calculation model . The former would be the case, for example, when determining a distance difference if the compared signals run through different layers of air or cannot be measured precisely simultaneously.
When taking measurements on the oscilloscope , a differential measurement of a quasi-difference z. B. differ by not completely suppressed common mode voltages of the two signals.
Similarly, quasi-differences in econometrics are a variant of dynamic DL models when special disturbance variables are effective. In the case of forecasts with quasi-differences, the question is whether a “soft” filter cannot be more suitable than a “hard” difference formation.
In the case of digital filters , the difference filter differs from the quasi-difference filter in that it has a transfer function that is usually applied exponentially.
In computing , the term quasi-difference is also used for calculating machines that work with numbers that are related to the values represented by them in a non-linear transformation (e.g. quasi-sum or quasi-difference of logarithms ).
Sources and web links
- Helmut Thome, time series analysis . Section 3.12, Models for Non-Stationary Time Series. Oldenbourg Wissenschaftsverlag, Munich 2005
- Parameter estimation in dynamic models with ADL form or quasi-differences (Wirtschaftsuniv.Wien) ( MS PowerPoint ; 277 kB)
- Spectral analysis of time series, gapped-weight filters
- Quasi-addition and subtraction of a logarithmic Zuse calculating machine ( Memento from February 12, 2013 in the web archive archive.today )