Check Scoring Results for Concepts and Free Terms

This section provides an overview of what you can find in the Test Extractor Dialogue's results regarding concepts free terms.

Concepts Scoring Results

The results of the Test Extraction Dialogue's settings in regard to concepts provide the following information table (1) ordered from highest to lowest score value, relating also to SKOS properties and definitions:

  • Concept Preferred Label: this column lists concepts found in the text sample. They are links you can click on and this opens the concept's Linked Data Frontend in PoolParty.

  • Normalized Score: the normalized score is a mathematical value to show relative relations of the Corpora Score values. The highest Corpora Score value thus will be always equaled to '100.0'. Other scores will appear here as a relative value to the highest. This makes it easier to compare score values and evaluate them.

  • Corpora Score: the values in this column show calculation results of the concepts in the list. They represent the mean of frequency and relevance, thus displaying how often a concept can be found in the corpus, individual documents as well as this text sample, and how relevant it is in respect to the thesaurus you chose in the settings. Specifically, PoolParty's extractor juxtaposes the number of documents, number of times a term or a concept is found and the text lengths.

  • Transitive Broaders: results in this column refer to the concepts found in relation to the one in the Concept Preferred Label column, they represent concepts that are found as transitive broader concepts, according to SKOS rules.

  • Transitive Broader Top Concepts: top concepts that are of the transitive property type will be listed here.

    • In both columns you can click the URI's to open the corresponding display in PoolParty's Linked Data Frontend.

  • Related Concepts: this column (empty in this example) displays concepts of the corresponding SKOS property.

You can also click Display Debug Information (2), which opens a new dialogue, details find here: Test Extraction Results - Debug Information Dialogue


Free Terms Scoring Results

The results for Free Terms in this view are similar to the ones for concepts. Free terms here are terms the extractor has found in the sample you used in the Text Extractor.

Additional values here are the following:

  • Term: a column which just lists the terms found. These terms are also based on the co-occurrence feature and thus represent terms that frequently occur in the text sample you used and additionally in close proximity to actual concepts.

  • Frequency in Document: this value represents an actual number of occurrences of this term in the sample text.


Test Extraction Results - Debug Information Dialogue

In this section you find details about the Debug Information Dialogue provided in the Text Extraction Results window.

The Test Extraction is based on the PoolParty extractor. The results listed in a tabular overview in the Text Extraction Results will be shown in code displays in the Debug Information Dialogue.

The following sections are available here:

  • Document: this section contains all concepts found in the text sample you provided. Additionally you find information on the extracted and stored data, such as the unique system's ID for that concept, the project ID and information about URIs, pertaining labels and SKOS relations.

  • Terms (2): scrolling down the page you find this section of the dialogue displaying terms that have been extracted from the text sample, and the data pertaining to them, such as their position in the text, the corpora score value of each and frequency or concept match information.

To leave the dialogue, use the Close icon (3) in the top right-hand corner, click Close in the bottom right-hand corner or click into a visible part of the Text Extraction Results window underneath.