files

II Edit

The Edit menu is available only in the advanced mode.

View data; it allows to display the data either as two-dimensional plot as well as landscapes in the different modes.
Permute; it allows to change the order of the modes in a single set data set
Reshape; it allows to reshape (also mixing the modes, if one desires) a single set data set (i.e. both X and Y)
Labels & Info; it allows to edit mode, sample and variables labels, description, author etc.
Select; it allows the selection of subsets of data for computing/validating a certain model

All data sets can be displayed as two dimensional plots as well as landscapes, provided that the data set has three dimensions or less.

As for the landscapes, the plot refers to a single matrix, in the case of two dimensional plots it is then the complete data matrix, whereas for three-way arrays it is necessary to fix one of the modes. The fixed mode is given by the submenu: "Mode 1", "Mode 2" or "Mode 3".
Having loaded the Fluorescence data set (both X and Y). And selecting: Edit -> View data -> Landscape -> Mode 3, the following requester appears:

It is indeed possible to choose only one variable.
If one selects "285 nm", the image on the right appears.

A two dimensional plot relative to one mode (for a three way array) displays one plot for each variable in the selected dimension:

If the first mode is selected then 15 plots (for the Fluorescence data set) are displayed.
Each of them has length 990 (that is 66 x 15).
In other words the complete data set is displayed with the two non fixed modes on the X axis. The chosen unfolding is then the one having the fixed mode as rows and the two others as columns.

For the picture on the left, this means that the first part (up to the first vertical line) contains the emission spectra at the first excitation wavelength.

The "philosophy" for two dimensional arrays is the same, therefore if one wants to display the columns (i.e. the variables) of a two way array the second mode is to be chosen

The figure here refers to the Y array for the Fluorescence data set.

Permute

"Single set" data sets can be permuted in any desireable fashion. The only limitation is when both X and Y are loaded; in this case any permutation is accepted so long as at least the first mode remains the same (in other terms, they must have the first mode in common) for both the arrays.

Upon loading the Fluorescence data set has dimensions: 15 x 66 x 15

By selecting Edit -> Permute The window on the left opens; if one wants to have "Emission" as the third mode he/she has to type 3 in the corresponding edit object and 2 in the edit object relative to the third mode.

After pressing OK on the "Permute data" window the InfoBox shows the new dimensions of the permuted data set.
If the same "mode-index" is repeated an error message is produced, the same occurs if the dimensions in the first mode are not identical for X and Y or if the labels in the first mode for X and Y are not the same.

Reshape

One or more modes can be split or joined by using this menu.
If both X and Y are present it is initially asked which one one desires to reshape:

The following requester allows to decide whether one wants to concatenate ("decrease") two or more modes or to split ("increase") them:

If one chooses "decrease" a window such as the following appears:

The checked mode is the one to be split; in this case (albeit for this specific data set the choice is meaningless), it is decided to split the first mode into a 5 x 3 array.
The "Evaluate" button will return an error message if the chosen reshaping is not feasible. The new dimensions will be visible immediately above the said button.
The "Default" button returns to the original dimensions so that a new reshape operation can be performed.

Press "OK" to accept the reshaping and continue.

If one chooses decrease two or more modes are joined into one (i.e. it is possible to matricise - unfold - an array)

First select the two modes to concatenate By pressing "Evaluate" the new dimensions are displayed.
In this case the array is matricised in the form I x JK.
After pressing OK the InfoBox shows the new dimensions for the array.

NB In case X and Y are present the same reshaping shall occurs on both if the first mode is interested: if after the reshaping the two first modes have different labels (or size in the first mode) an error message will be displayed.

Labels & Info

This menus allows to edit the additional information stored associated to the data (both X and Y).
It has one or two submenus depending on whether a Y matrix/array is available.

If one imports the array "X" from the "claus.mat" data set, standard labels are created (i.e. "Sam. #" in the first mode and "Var. #" in the others). The labels editor appears then as:

In the upper right corner the "name" of the data set is visible, immediately underneath are the dimensions.
The listbox in the left part specifies the current labels in the mode defined in the "Mode" frame.
There are different ways of defining the labels in a certain mode.
It is possible to import them from a .mat file or from a .txt file.
In the first case there are three possibilities:

Cell vector of chars with a number of elements equal to the number of selected variables

Char array with a number of rows equal to the the number of selected variables

Vector of scalars with length equal to the number of variables in the mode (this option is not active unless or variables were selected). The same values will be employed as scalars to plot against in the models as well as when displaying the raw data.

In case one wants to import the labels from a text file, the latter must have the label for one variable on each line, i.e. the end of the line is considered as the end of the label.

For instance, to import from "claus.mat" the emission wavelengths:

Change the mode to "Mode 2" By right-clicking on the listbox a context-menu appears.
Choose select all.

Select, either from the context menu or from the menu "Files", the "import -> from .mat" menu.
A requester appears asking for the file and subsequently (unless the file contain only one variable) which variable to load.

Select EmAx To define the labels a measurement unit is asked. In this case insert "nm".

NB The measurement unit can also be empty; in such case the labels will be identical to the loaded scalars.
The new labels are finally visible in the list box.

The name of the data set can be changed via the: Edit -> Rename data set menu;

If one wants to change the information relative to the current mode (e.g. in the claus data set to rename the second mode as "Emission"), he/she can choose Edit -> Rename Mode
The requester on the right appears:

"Name" is the name appearing in the menus as well as in the titles whenever there is reference to the specific mode.
"Axis labels" appears under the axes whenever something (e.g. the loadings) are plot versus the scalars in this mode.
"Default labels" are used when the menu Edit -> Default Names is chosen. The selected labels are then in the form:
"Default-Label #1".

It is possible to rename the single variables also within the editor: select some variables in the listbox and choose the Edit ->Rename Variables menu (only Rename in the context menu). A requester allowing to edit the selected labels appears.

NB The renaming of the variables does not affect the scalars (even when the scalars have been used to define the labels themselves).

One last possibility is to use the "Define variables as -> Prefix + Number" or "Define variables as -> Number + Meas Un." menus (the second is available only if all the variables in the current mode have been selected). A suitable requester appears:

After pressing on the OK button the labels are updated

NB The numerical values (only when defined as the Number + Meas. Unit) are also used as scalars in the plots.

If one wants to return to the default values he/she shall just select "Default names" in the Edit menu or in the context menu.

The change will affect only the selected variables:

It is also possible to edit Author and Description associated to the data set via the Edit -> Author and Edit -> Description menus. A suitable requester appears:

By means of the View -> Author and View -> Description menus

it is also possible to display these to information:

Select

It makes possible to apply the desired models to subsets of samples/batches or variables; the choice of which one to use is made via the Edit -> Select submenus:

NB If only X is available the Select Y Variables is not displayed.

With Select Batches one can assign the Samples/Batches to the test set/calibration set:

The asterisks identify the samples/batches that belong to the calibration set.

If a T is also present the sample/batch belongs to the Test set (this option is relevant only for those models - such as (n)PLS1 - that support the test set validation).
In the figure on the left sample 6 and 14 are left out of the computation while samples 4 and 10 represent the calibration set.

In the "Groups" list box the numbers in parenthesis are respectively the number of samples/batches in the calibration set and the number of samples/batches in the test set.

A context menu in the "Samples" list box allows to highlight subsets in the sample/batch mode

or to remove the highlighted samples/batches from both calibration and test set (remove all from both).

The "Select X Variables" (or Select Y Variables) menu, allows to select subsets of variables in the 2nd, 3rd, ... mode.

The 'X *' (or 'Y *') marker identifies the variables that will be employed in the computation.

The numbers in parenthesis beside the Modes' names are the variables assigned to the predictors (the first) or the predictands (the second). It is currently possible to assign X variables to the predictors only and Y variables to the predictands only.

If no variables are selected in one of the modes but there are selections in others, the complete mode will be used.

A context menu allows to select subsets of variables (one every ... of the highlighted ones), to select them all or to desect the highlighted ones.

Change the mode to "Mode 2"	By right-clicking on the listbox a context-menu appears. Choose select all.	Select, either from the context menu or from the menu "Files", the "import -> from .mat" menu. A requester appears asking for the file and subsequently (unless the file contain only one variable) which variable to load.
Select EmAx	To define the labels a measurement unit is asked. In this case insert "nm". NB The measurement unit can also be empty; in such case the labels will be identical to the loaded scalars.	The new labels are finally visible in the list box.