calcLRs.YsByXs: Calculate linear regressions for time series in Y as function...

In wStockhausen/wtsDisMELSConn: Functions for dealing with DisMELS IBM model output, including calculating and analyzing connectivity matrices

Description

Function to calculate linear regressions between time series in Y as function of time series in X.

Usage

calcLRs.YsByXs(mdfrX, mdfrY, verbose = FALSE)

Arguments

mdfrX	- melted dataframe with independent (X) time series
mdfrY	- melted dataframe with dependent (Y) time series
verbose	- flag to print messages

Details

Requires packages 'plyr', 'reshape2', and 'glmulti'. Datasets are standardized as z-scores, so the constant term in the linear model is dropped.

none.

Value

list with named elements 'res', 'summary' and 'plot':

• res - a list with sublists by unique YG:YV:XG:XV with elements

  • lms - list with linear model (lm) results by Y for X

  • summary - dataframe with summaries of linear model results

• summary - a dataframe summarizing all LR results. Columns are

  • ygroup - dependent variable group label

  • y - dependent variable label

  • xgroup - independent variable group label

  • x - independent variable label

  • n - number of valid data points

  • rho - Pearson's correlation coefficient

  • rsq - R^2 for linear fit

  • adj.rsq - adjusted R^2 for linear fit

  • F - F statistic

  • prF - Pr(>F) (p-value uncorrected for multiple comparisons)

  • aicc - AICc value

  • del.aicc - AICc relative to null model [i.e., AICc(model) - AICc(null)]

• data - dataframe with data involved in linear models

  • ygroup - dependent variable group label

  • yvar - dependent variable label

  • xgroup - independent variable group label

  • xvar - independent variable label

  • year - year

  • x - independent value in LR

  • y - dependent value in LR

wStockhausen/wtsDisMELSConn documentation built on May 3, 2019, 7:36 p.m.