Themes#
Application ID
eot_themes
Application Aliases
themes.app eot/themes.app
Description#
The Themes application discovers the most relevant themes or categories in your documents and the words that have been identified as clues for each enclosed theme.
Configuration#
The configuration is an object defined as (bold = required, italic = optional):
collect (
array[string]
) – The array of concepts of interestcollect:uri (
boolean
) – A boolean flag indicating if the concept should be includedcollect:exclude (
object
) – An object with the attributes, key values to be excludedattributes (
array[string]
) – The attributes used for the theme categoriescount (
number
) [optional
] – The maximal number of topics to be returned. Must be strictly positivethreshold (
number
) [optional
] – The minimal threshold value of the topics. Valid values are in the range[0,100]
debug_info (
boolean
) – Add debug information in the results.
For example, to collect the Person
, Company
and Event
where:
Person: we do not include the URI, but we look at the
"position"
attributeCompany: we keep the URI but we exclude all companies in the
"media"
sector, as they are often mentioned in articles but do not really add valueEvent: we take the attributes
"sector"
and"theme"
the following configuration would be used:
{
"collect": {
"Person": { "uri": false, "attributes": ["position"]},
"Company": { "uri": true, "exclude": {"sector": ["media"]}},
"Event": {}
},
"attributes": ["sector", "theme"]
}
Example#
Identifying the most prominent themes:
wow -p 'english,semantic-theme,themes.app' \
-i "There have been 780 confirmed cases of monkeypox, the World Health Organization says."
which yields:
[
{
"name": "healthcare",
"relevancy": 100
},
{
"name": "viral infection",
"relevancy": 67
},
{
"name": "pathology",
"relevancy": 33
}
]
For an interpretation of the JSON data, refer to the application’s JSON schema.