Mean Fuser
==========

.. figure:: icons/mean-fuser.png

Constructs relation matrices based on the average values of matrix
elements.

Signals
-------

**Inputs**:

-  **Fusion Graph**

   A relational scheme of a data compendium.

-  **Relation**

   Relationships between two groups of objects.

**Outputs**:

-  **Mean-fitted fusion graph**

   Mean fuser.

-  **Relation**

   Relationships between two groups of objects.

Description
-----------

The widget completes each relation matrix at the input based on the
available data in the matrix. Unknown values in the matrix can be
replaced with the values obtained by averaging matrix rows, matrix
columns or the entire data matrix.

.. figure:: images/MeanFuser-stamped.png

1. Select the axis for mean value calculation:

   -  **rows**
   -  **columns**
   -  **all**
   
2. Output selected relation matrix, where unknown matrix elements are
   replaced with mean values.

Example
-------

**Mean Fuser** widget is useful for comparing RMSE values in
**Completion Scoring** widget for the input data set. In the example
below we have sampled movie ratings, fed the in-sample movie ratings
data into **Fusion Graph** and from there into **Completion Scoring**
for evaluation. We also fed the out-of-sample data from **Matrix
Sampler** into **Completion Scoring** widget as out-of-sample movie
ratings data is needed to assess how well the predicted values
correspond to the true data. Finally, we compare prediction to those
obtained by **Mean Fuser**.

.. figure:: images/MeanFuser-Example.png