Dictionaries
Lingo3G uses dictionaries to improve the quality of clustering for a specific language. This article shows how to customize dictionaries in the REST API.
Customizing global dictionaries
To customize the global dictionaries for REST API clustering, go to the DCS dictionaries folder and edit the files as required. Once you finish editing, restart the DCS for the changes to take effect.
Using per-request dictionaries
You can provide extra per-request dictionary entries for a specific clustering request. Lingo3G applies these extra entries as an addition to the default dictionaries. For example, if the end-user wants to remove specific labels from the clustering result they are currently viewing, your software can add such labels to the per-request label dictionary and rerun the clustering.
The dictionaries section of the Lingo3G parameters object groups all per-request dictionaries. You can use any combination of per-request label, synonym and tag dictionaries.
Label dictionaries
To add per-request entries to the
label dictionary, use the
dictionaries.labels
parameter.
In the following example, without the per-request label dictionary, Lingo3G creates a cluster called Configuration Issue on Windows. Adding two entries to the per-request dictionary replaces the label with Windows:
Synonym dictionaries
To add per-request entries to the
synonym dictionary, use the
dictionaries.synonyms
parameter.
In the following example, adding Chrome and Firefox as synonyms puts documents containing either of these words in the same cluster.
Note that setting theclusters.allowOneDocumentClusters
to true
is required due to the small number of documents in
the example. The parameter causes Lingo3G to preserve during processing
words with one occurrence, such as Firefox and
Chrome in the example. If the example left
allowOneDocumentClusters
at its default value of false
, documents containing
Chrome an Firefox
would not end up in the same cluster because Lingo3G would filter out
the one-occurrence words before applying synonyms.
Tag dictionaries
To add per-request entries to the
word tag dictionary, use the
dictionaries.tags
parameter.
In the following example, adding Chrome and Firefox to
the tags dictionary with the name
tag slightly promotes the
two words in cluster labels. As a result, Lingo3G chooses
Chrome and Firefox over some other words to label the
clusters.