Generalized R-squared for detecting dependence
نویسندگان
چکیده
Detecting dependence between two random variables is a fundamental problem. Although the Pearson correlation coefficient is effective for capturing linear dependence, it can be entirely powerless for detecting nonlinear and/or heteroscedastic patterns. We introduce a new measure, G-squared, to test whether two univariate random variables are independent and to measure the strength of their relationship. The G-squared statistic is almost identical to the square of the Pearson correlation coefficient, R-squared, for linear relationships with constant error variance, and has the intuitive meaning of the piecewise R-squared between the variables. It is particularly effective in handling nonlinearity and heteroscedastic errors. We propose two estimators of G-squared and show their consistency. Simulations demonstrate that G-squared estimators are among the most powerful test statistics compared with several state-of-the-art methods.
منابع مشابه
The dependence of the modulus of elasticity on reference temperature in the theory of generalized thermoelastic diffiusion with one relaxation time
متن کامل
Detecting the location of the boundary layers in singular perturbation problems with general linear non-local boundary conditions
Singular perturbation problems have been studied by many mathematicians. Since the approximate solutions of these problems are as the sum of internal solution (boundary layer area) and external ones, the formation or non-formation of boundary layers should be specified. This paper, investigates this issue for a singular perturbation problem including a first order differential equation with gen...
متن کاملTruncated Linear Minimax Estimator of a Power of the Scale Parameter in a Lower- Bounded Parameter Space
Minimax estimation problems with restricted parameter space reached increasing interest within the last two decades Some authors derived minimax and admissible estimators of bounded parameters under squared error loss and scale invariant squared error loss In some truncated estimation problems the most natural estimator to be considered is the truncated version of a classic...
متن کاملAn Evaluation of an Adaptive Generalized Likelihood Ratio Charts for Monitoring the Process Mean
When the objective is quick detection both small and large shifts in the process mean with normal distribution, the generalized likelihood ratio (GLR) control charts have better performance as compared to other control charts. Only the fixed parameters are used in Reynolds and Lou’s presented charts. According to the studies, using variable parameters, detect process shifts faster than fixed pa...
متن کاملSome numerical studies of the evolution of generalized parton distributions
We study the evolution behavior of generalized parton distributions at small longitudinal momentum fraction. Particular attention is paid to the ratio of a generalized parton distribution and its forward limit, to the mixing between quarks and gluons, and to the dependence on the squared momentum transfer t.
متن کامل