|
|
New Models for Automatic Detection of Performance Degradation
Stupinský, Šimon ; Češka, Milan (referee) ; Rogalewicz, Adam (advisor)
Performance testing is a critical factor in the optimisation of programs during its development, but it is still not so well developed in comparison to functional testing. A framework Perun provides full automation of performance management, thereby contributing to the development of this area. We have introduced three non-parametric approaches to performance data modelling: regressogram, moving average and kernel regression, which were integrated within this framework. We try to achieve appropriate approximations of performance data using these techniques, without the assumption of dependence between two variables, which represents the main advantage in comparison to parametric techniques. Further, we have proposed and implemented two methods for automatic detection of performance changes, which works with all kinds of models within Perun . We have demonstrated our solutions on the real project ( Vim ), and on the set of the experimental cases, in which we compared proposed solutions with existing. We have achieved decreased time processing about two-thirds and an almost triple improvement in the fitness of data modelling with new modelling approaches. The proposed detection methods detected performance degradation of three specific functions in comparison of two different versions of Vim, where was present a known performance issue.
|
|
|
New Models for Automatic Detection of Performance Degradation
Stupinský, Šimon ; Češka, Milan (referee) ; Rogalewicz, Adam (advisor)
Performance testing is a critical factor in the optimisation of programs during its development, but it is still not so well developed in comparison to functional testing. A framework Perun provides full automation of performance management, thereby contributing to the development of this area. We have introduced three non-parametric approaches to performance data modelling: regressogram, moving average and kernel regression, which were integrated within this framework. We try to achieve appropriate approximations of performance data using these techniques, without the assumption of dependence between two variables, which represents the main advantage in comparison to parametric techniques. Further, we have proposed and implemented two methods for automatic detection of performance changes, which works with all kinds of models within Perun . We have demonstrated our solutions on the real project ( Vim ), and on the set of the experimental cases, in which we compared proposed solutions with existing. We have achieved decreased time processing about two-thirds and an almost triple improvement in the fitness of data modelling with new modelling approaches. The proposed detection methods detected performance degradation of three specific functions in comparison of two different versions of Vim, where was present a known performance issue.
|
|
|
Varying coefficient models
Sekera, Michal ; Maciak, Matúš (advisor) ; Komárek, Arnošt (referee)
The aim of this thesis is to provide an overview of the varying coefficient mod- els - a class of regression models that allow the coefficients to vary as functions of random variables. This concept is described for independent samples, longi- tudinal data, and time series. Estimation methods include polynomial spline, smoothing spline, and local polynomial methods for models of a linear form and local maximum likelihood method for models of a generalized linear form. The statistical properties focus on the consistency and asymptotical distribution of the estimators. The numerical study compares the finite sample performance of the estimators of coefficient functions. 1
|
guest ::
