likelihood

Return log-likelihood function (cost function) value evaluated at a given parameter tensor, an observed tensor, and cut-off points.

Efficient algorithms for tensor noise reduction and completion. This package includes a suite of parametric and nonparametric tools for estimating tensor signals from noisy, possibly incomplete observations. The methods allow a broad range of data types, including continuous, binary, and ordinal-valued tensor entries. The algorithms employ the alternating optimization. The detailed algorithm description can be found in the following three references.

Chanwoo Lee

TensorComplete

Tensor Noise Reduction and Completion Methods

likelihood function

<dl><dt>ttnsr</dt>
<dd>An observed tensor data.</dd>
<dt>theta</dt>
<dd>A continuous-valued tensor (latent parameters).</dd>
<dt>omega</dt>
<dd>The cut-off points.</dd>
<dt>type</dt>
<dd>Types of log-likelihood function.
<code>"ordinal"</code> specifies log-likelihood function based on the cumulative logistic model.
<code>"Gaussian"</code> specifies log-likelihood function based on the Gaussian model.</dd></dl>

Arguments

Log-likelihood function (cost function). — likelihood

<dl>

<dt>ttnsr</dt>
<dd>An observed tensor data.</dd>


<dt>theta</dt>
<dd>A continuous-valued tensor (latent parameters).</dd>


<dt>omega</dt>
<dd>The cut-off points.</dd>


<dt>type</dt>
<dd>Types of log-likelihood function.
<code>"ordinal"</code> specifies log-likelihood function based on the cumulative logistic model.
<code>"Gaussian"</code> specifies log-likelihood function based on the Gaussian model.</dd>

</dl>

likelihood: Log-likelihood function (cost function).

Description

Usage

Value

Arguments