POMEN TESTIRANJA HIPOTEZ V DEFORMACIJSKI ANALIZI
THE SIGNIFICANCE OF HYPOTHESIS TESTING IN DEFORMATION ANALYSIS

Simona Savšek

DOI: 10.15292/geodetski-vestnik.2013.03.465-478

 

Izvleček:

V geodeziji na podlagi kakovostnih opazovanj ter ustreznega matematičnega in stohastičnega modela pridobimo zanesljive rezultate. Izpolnjevanje nekaterih meril in zahtev pri merskih in računskih postopkih ugotavljamo z ocenjevanjem in testiranjem značilnih parametrov, pri čemer so postopki statističnega testiranja hipotez nepogrešljivi. V deformacijski analizi s statističnimi testi preizkušamo domneve o parametrih, ugotavljamo skladnost opazovanj in matematičnega modela, odkrivamo grobe pogreške v opazovanjih ter ugotavljamo skladnost domnevno stabilnih točk med terminskimi izmerami in določamo točke s statistično značilnimi premiki. Naloga geodeta je, da za obdelavo kakovostnih merskih podatkov uporabi pravilen in zanesljiv matematični model, ki zagotavlja visoko stopnjo zaupanja v rezultate in tako zmanjšuje naročnikovo tveganje. Postopki statističnega testiranja hipotez pomagajo pri odločitvah, ne morejo pa nadomestiti merskih in računskih postopkov v geodeziji.

Ključne besede: statistično testiranje hipotez, stopnja značilnosti testa, porazdelitvena funkcija, kritična vrednost, dejansko tveganje

 

Abstract:

In geodesy, reliable results are obtained with quality observations and the corresponding mathematical and stochastic models. One establishes if individual criteria and requirements during the measurement and calculation procedures are met by evaluating and testing typical parameters respectively – procedures of statistical hypothesis testing being indispensable part of the process. In deformation analysis, statistical tests are applied for testing the assumptions about the parameters, for assessing the conformity of observations and the mathematical model, for detecting of gross errors in observations, for identifying the conformity of allegedly fixed points between the
epochs and for determining the points with statistically typical displacements. For the processing of quality measurement data, the task of a surveyor is to use a correct and reliable mathematical model, which provides a high degree of confidence in the results and reduces a client's risk. Methods of statistical hypothesis testing help experts to make decisions, but it is not possible to use them to replace measurement and calculation procedures in geodesy.

Keywords: statistical hypothesis testing, significance level, distribution function, critical value, actual risk

 

Literatura / References:

Ambrožič, T., Turk, G., Stopar, B. (2002). Premik – računalniški program za izračun in ugotavljanje statistično značilnih premikov v položajni geodetski mreži. Interna izdaja, Ljubljana.

Caspary, W. F. (2000). Concepts of Network and Deformation Analysis. School of Surveying. Kensington: University of New South Wales.

Grigillo, D., Stopar, B. (2003). Metode odkrivanja grobih pogreškov v geodetskih opazovanjih. Geodetski vestnik, 47(4), 387–403.

Harvey, B. R. (1991). Practical Least Squares and Statistics For Surveyors, University of New Outh Wales, School of Surveying, Kensington.

Kuang, S. (1969). Geodetic Network Analysis and Optimal Design: Concepts and Applications, Ann Arbor Press, Inc., Chelsea, Michigan.

Savšek-Safić, S. (2002). Optimalna metoda določanja stabilnih točk v deformacijski analizi. Doktorska disertacija. Ljubljana: FGG OGG.

Savšek-Safić, S., Ambrožič T., Stopar B., Turk G. (2006). Determinantion of Point Displacements in the Geodetic Network. Journal of Surveying Engineering, 132(2), 58–63.