Scorer (JavaScript)

Compares two columns by their attribute value pairs and shows the confusion matrix, i.e. how many rows of a given attribute and its classification match. The dialog allows you to select two columns for comparison; the values from the first selected column are represented in the confusion matrix's rows and the values from the second column by the confusion matrix's columns. The view of the node displays three tables, the first one is the confusion matrix with the number of matches in each cell. Row and column rates can be shown via a configuration setting; they are the number of correct predictions divided by the total number of records in the row/column. Additionally, it is possible to highlight cells of this matrix to select the underlying rows. The selection can be passed to other JavaScript views. The second table reports a number of statistics specific to a given class such as True-Positives, False-Positives, True-Negatives, False-Negatives, Accuracy, Balanced Accuracy, Error Rate, False Negative Rate, Recall, Precision, Sensitivity, Specificity, F-measure. The last table contains overall statistics like the Overall Accuracy, Overall Error, Cohen's kappa, the Correctly Classified and the Incorrectly Classified values. These three tables are also available as output ports of this node.

The node supports custom CSS styling. You can simply put CSS rules into a single string and set it as a flow variable 'customCSS' in the node configuration dialog. You will find the list of available classes and their description on our documentation page.

Options

Scorer Options

Title
The view title
Subtitle
The view subtitle
Actual column
The actual column contains the actual classes of the data.
Predicted column
The predicted column contains the predicted classes of the data.
Sorting Strategy
Whether to sort the labels according to their appearance, or use the lexical/numeric ordering.
Reverse order
Reverse the order of the elements.
Ignore missing values
Choose how to treat missing values in either the reference or prediction column. Default is to ignore them (treat them as if the row did not exist). Alternatively, you can expect the table to not contain missing values in these two columns. If they do, the node will fail during execution.
Header color
Background color used for the header sections of the tables (rows or column headers).
Diagonal color
Background color used for highlighting the values on the diagonal of the confusion matrix (the correct predictions).
Display table headers
Check to display the "confusion matrix", "class statistics" and "overall statistics" labels.
Display class nature
Check to display the class nature (Actual or Predicted).
Display number of rows
Check to display the total number of rows in the top left corner of the confusion matrix.
Display float values as percentages
Check to display non-integer values as percentages.
Display confusion matrix rates
Check to display the confusion matrix's rows and columns totals as rates. These are the result of the division of the number of correct predictions on a given row (or cell) divided by the sum of all values on that row or column.
Display fullscreen button
Check to display a button which switches the view into fullscreen mode. The button is only available in the KNIME WebPortal.
Show warnings in view
If checked, warning messages will be displayed in the view when they occur.

Statistics Options

Display class statistics table
Check to display the class statistics table.
Calculate/display true positives for each class
Check to display the true positives in the class statistics table.
Calculate/display false positives for each class
Check to display the false positives in the class statistics table.
Calculate/display true negatives for each class
Check to display the true negatives in the class statistics table.
Calculate/display false negatives for each class
Check to display the false negatives in the class statistics table.
Calculate/display accuracy for each class
Check to display the accuracy in the class statistics table.
Calculate/display balanced accuracy for each class
Check to display the balanced accuracy in the class statistics table.
Calculate/display error rate for each class
Check to display the error rate in the class statistics table.
Calculate/display false negative rate for each class
Check to display the false negative rate in the class statistics table.
Calculate/display recall for each class
Check to display the recall in the class statistics table.
Calculate/display precision for each class
Check to display the precision in the class statistics table.
Calculate/display sensitivity for each class
Check to display the sensitivity in the class statistics table.
Calculate/display specificity rate for each class
Check to display the specificity in the class statistics table.
Calculate/display F-measure for each class
Check to display the F-measure in the class statistics table.
Display overall statistics table
Check to display the overall statistics table.
Calculate/display overall accuracy
Check to display the overall accuracy in the overall statistics table.
Calculate/display overall error
Check to display the overall error in the overall statistics table.
Calculate/display Cohen's kappa
Check to display the Cohen's kappa in the overall statistics table.
Calculate/display number of correctly classified values
Check to display the number of correctly classified values in the overall statistics table.
Calculate/display number of incorrectly classified values
Check to display the number of incorrectly classified values in the overall statistics table.

Control Options

Enable view edit controls
Check to enable controls of the interactive menu in the chart.
Enable title edit controls
Check to enable the editing of the title within the view.
Enable subtitle edit controls
Check to enable the editing of the subtitle within the view.
Enable display table headers toggle
Check to enable the toggling of the display of the table headers.
Enable display rows number toggle
Check to enable the toggling of the display of the rows number.
Enable display confusion matrix rates toggle
Check to enable the toggling of the display of the Confusion Matrix's rows and columns totals as rates.
Enable display class statistics table toggle
Check to enable the toggling of the display of the class statistics table.
Enable display overall statistics table toggle
Check to enable the toggling of the display of the overall statistics table.

Input Ports

Icon
Table containing at least two columns to compare.

Output Ports

Icon
The confusion matrix
Icon
The class statistics table
Icon
The overall statistics table

Views

Interactive View: Confusion Matrix

Workflows

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.