Generalized Correlations and Initial Causal Path

Since causal paths from data are important for all sciences, the package provides sophisticated functions. The idea is simply that if X causes Y (path: X to Y) then non-deterministic variation in X is more "original or independent" than similar variation in Y. We compare two flipped kernel regressions: X=f(Y, Z) and Y=g(X,Z), where Z are control variables. Our first two criteria compare absolute cross products of regressor values and residuals (Cr1) and absolute residuals (Cr2), are both quantified by stochastic dominance of four orders (SD1 to SD4). Our third criterion (Cr3) expects X to be better able to predict Y than the other way around using generalized partial correlation If |r*(x|y)|> |r*(y|x)| it suggests that y is more likely the "kernel cause" of x. The usual partial correlations are generalized with a new non-symmetric matrix developed here. Partial correlations help asses effect of x on y after removing the effect of a set of variables. The package provides additional tools for causal assessment, for printing the causal directions in a clear, comprehensive compact summary form, for matrix algebra, for "outlier detection", and for numerical integration by the trapezoidal rule, stochastic dominance, etc. The package has functions for bootstrap-based statistical inference and one for a heuristic t-test.

10,388

436

48

15

- Package
- generalCorr
- Type
- Package
- Title
- Generalized Correlations and Initial Causal Path
- Version
- 1.1.1
- Date
- 2018-01-24
- Author
- Prof. H. D. Vinod, Fordham University, NY.
- Maintainer
- H. D. Vinod
- Encoding
- UTF-8
- Depends
- R (>= 3.0.0), np (>= 0.60), xtable (>= 1.8), meboot (>= 1.4), psych (>= 1.5)
- Suggests
- R.rsp
- VignetteBuilder
- R.rsp
- Description
- Since causal paths from data are important for all sciences, the package provides sophisticated functions. The idea is simply that if X causes Y (path: X to Y) then non-deterministic variation in X is more "original or independent" than similar variation in Y. We compare two flipped kernel regressions: X=f(Y, Z) and Y=g(X,Z), where Z are control variables. Our first two criteria compare absolute cross products of regressor values and residuals (Cr1) and absolute residuals (Cr2), are both quantified by stochastic dominance of four orders (SD1 to SD4). Our third criterion (Cr3) expects X to be better able to predict Y than the other way around using generalized partial correlation If |r*(x|y)|> |r*(y|x)| it suggests that y is more likely the "kernel cause" of x. The usual partial correlations are generalized with a new non-symmetric matrix developed here. Partial correlations help asses effect of x on y after removing the effect of a set of variables. The package provides additional tools for causal assessment, for printing the causal directions in a clear, comprehensive compact summary form, for matrix algebra, for "outlier detection", and for numerical integration by the trapezoidal rule, stochastic dominance, etc. The package has functions for bootstrap-based statistical inference and one for a heuristic t-test.
- License
- GPL (>= 2)
- LazyData
- true
- RoxygenNote
- 6.0.1
- NeedsCompilation
- no
- Packaged
- 2018-01-24 18:31:25 UTC; hd
- Repository
- CRAN
- Date/Publication
- 2018-01-24 19:03:37 UTC