lc_cov_1d

Local constant covariance estimation

A general spatiotemporal satellite image imputation method based on sparse functional data analytic techniques. The imputation method applies and extends the Functional Principal Analysis by Conditional Estimation (PACE). The underlying idea for the proposed procedure is to impute a missing pixel by borrowing information from temporally and spatially contiguous pixels based on the best linear unbiased prediction.

Weicheng Zhu

stfit

Spatio-Temporal Functional Imputation Tool

lc_cov_1d function

<dl><dt>ids</dt>
<dd>a vector indicating subject/group ids</dd>
<dt>time</dt>
<dd>integer vector of observed time points, the minimum time unit is 1</dd>
<dt>resid</dt>
<dd>vector of residual values used for covariance calculation</dd>
<dt>W</dt>
<dd>weight vector, it contains both kernel and bandwidth information in general 
local polynomial estimation setting up</dd>
<dt>t1</dt>
<dd>time point 1</dd>
<dt>t2</dt>
<dd>time point 2</dd></dl>

Arguments

Local constant covariance estimation — lc_cov_1d

<dl>

<dt>ids</dt>
<dd>a vector indicating subject/group ids</dd>


<dt>time</dt>
<dd>integer vector of observed time points, the minimum time unit is 1</dd>


<dt>resid</dt>
<dd>vector of residual values used for covariance calculation</dd>


<dt>W</dt>
<dd>weight vector, it contains both kernel and bandwidth information in general 
local polynomial estimation setting up</dd>


<dt>t1</dt>
<dd>time point 1</dd>


<dt>t2</dt>
<dd>time point 2</dd>

</dl>

lc_cov_1d: Local constant covariance estimation

Description

Usage

Arguments