# Hst.sumup: Create Covariance Matrix In widals: Weighting by Inverse Distance with Adaptive Least Squares for Massive Space-Time Data

## Description

Calculate the covariance matrix of all model covariates

## Usage

 `1` ```Hst.sumup(Hst.ls, Hs = NULL, Ht = NULL) ```

## Arguments

 `Hst.ls` Space-time covariates. A list of length τ, each element containing a n x p_st numeric matrix. `Hs` Spacial covariates. An n x p_s numeric matrix. `Ht` Temporal covariates. An τ x p_t numeric matrix.

## Details

Important: The order of the arguments in this function is NOT the same as in the returned covariance matrix. The order in the covariance matrix is the same as in other functions in this package: `Hs`, `Ht`, `Hst.ls`.

## Value

A (p_s+p_t+p_st) x (p_s+p_t+p_st) numeric, symmetrix, non-negative definite matrix.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48``` ``` tau <- 20 n <- 10 Ht <- cbind(sin(1:tau), cos(1:tau)) Hs <- cbind(rnorm(10), rnorm(n, 5, 49)) Hst.ls <- list() for(tt in 1:tau) { Hst.ls[[tt]] <- cbind(rnorm(n, 1, 0.1), rnorm(n, -200, 21)) } Hst.sumup(Hst.ls, Hs, Ht) ########### standardize all covariates x1 <- stnd.Hst.ls(Hst.ls, NULL)\$sHst.ls x2 <- stnd.Hs(Hs, NULL, FALSE)\$sHs x3 <- stnd.Ht(Ht, n) Hst.sumup(x1, x2, x3) ## The function is currently defined as function (Hst.ls, Hs = NULL, Ht = NULL) { tau <- length(Hst.ls) if(tau < 1) { tau <- nrow(Ht) } if(is.null(tau)) { tau <- 10 ; cat("tau assumed to be 10.", "\n") } n <- nrow(Hst.ls[[1]]) if(is.null(n)) { n <- nrow(Hs) } big.sum <- 0 for (i in 1:tau) { if (!is.null(Ht)) { Ht.mx <- matrix(Ht[i, ], n, ncol(Ht), byrow = TRUE) } else { Ht.mx <- NULL } big.sum <- big.sum + crossprod(cbind(Hs, Ht.mx, Hst.ls[[i]])) } return(big.sum) } ```

widals documentation built on May 29, 2017, 10:11 p.m.