SalienceAttribute

A wrapper around the underlying SalienceAttribute structure in the C API.

`sAttribute`	The text of the attribute
`nCount`	The count of occurrences of the attribute
`nPositiveCount`	The count of positive occurrences of the attribute, based on value of the Neutral Upper Bound option
`nNegativeCount`	The count of negative occurrences of the attribute, based on value of the Neutral Lower Bound option
`nNeutralCount`	The count of neutral occurrences of the attribute, based on value of the Neutral Upper Bound and Neutral Lower Bound options
`Mentions`	A List<> of SaliencePhrase structures with details about the occurrences
`PositiveMentions`	A List<> of SaliencePhrase structures with details about the positive occurrences of the attribute, based on value of the Neutral Upper Bound option
`NegativeMentions`	A List<> of SaliencePhrase structures with details about the negative occurrences of the attribute, based on value of the Neutral Lower Bound option
`NeutralMentions`	A List<> of SaliencePhrase structures with details about the neutral occurrences of the attribute, based on value of the Neutral Upper Bound and Neutral Lower Bound options

SalienceChunk

A wrapper around the underlying SalienceChunk structure in the C API.

`sLabel`	A label describing the grammatical type of the chunk
`nSentence`	A zero-based index indicating the sentence in which the chunk appears
`Tokens`	A List<> of SalienceWord structures that comprise the chunk

SalienceCollectionDetails

A wrapper around the underlying SalienceCollectionDetails structure in the C API.

`nSize`	The number of documents added to the collection

SalienceCollectionDocument

`sIdentifier`	An identifier for the document
`sText`	The text of the document

SalienceCollectionEntity

A wrapper around the underlying SalienceCollectionEntity structure in the C API.

`sNormalizedForm`	Normalized form of the entity
`sType`	Type of entity (Company, Person, Place, Product, etc.)
`sLabel`	Descriptive label for the entity
`nCount`	Count of documents in which this entity occurs
`nPositiveCount`	Count of documents in which this entity occurs with positive sentiment, based on value of the Neutral Upper Bound option
`nNegativeCount`	Count of documents in which this entity occurs with negative sentiment, based on value of the Neutral Lower Bound option
`nNeutralCount`	Count of documents in which this entity occurs with neutral sentiment, based on value of the Neutral Lower Bound and Neutral Upper Bound options
`oMentions`	A List<> of SaliencePhrase structures of all mentions of this entity in the document

SalienceDocument

A wrapper around the underlying SalienceDocument structure in the C API.

`pSentences`	A List<> of SalienceSentence structures that make up the document

SalienceDocumentDetails

A wrapper around the underlying SalienceDocumentDetails structure in the C API.

`Sections`	A List<> of SalienceSection structures containing details about the document content

SalienceEntity

A wrapper around the underlying SalienceEntity structure in the C API.

`sNormalizedForm`	Normalized form of the entity
`sType`	Type of entity (Company, Person, Place, Product, etc.)
`sLabel`	Descriptive label for the entity
`fSentimentScore`	Sentiment related to the entity
`nEvidence`	A measure of how much evidence the sentiment score was based on (1 to 7)
`nConfident`	Whether the entity passed any confidence queries
`nAbout`	Whether the document is about this entity
`sSummary`	A summary of the document related to this entity
`nCount`	The number of entity mentions for this entity
`nFirstPos`	The byte offset within the content indicating the location of the first mention of the entity
`Mentions`	A List<> of SalienceMention structures of all mentions of this entity in the document
`Themes`	A List<> of SalienceTheme structures of themes related to this entity
`SentimentPhrases`	A List<> of SalienceSentimentPhrase structures of sentiment phrases related to this entity
`RelatedEntities`	A Dictionary<String,List<> > of SalienceEntity structures providing labeled references to entities related to this entity (currently used by meta-entities, likely to be expanded).
`Topics`	A List<> of SalienceTopic structures providing information about topics related to the entity

SalienceEntityParams

`nRequiredThreshold`	Threshold for entity extraction, sets the Entity Threshold option. Default value is 55.
`nSummaryLength`	Length in sentences for entity summary, sets the Entity Summary Length option. Default value is 2.
`nTimeout`	A timeout for entity extraction. Default value is 5.

SalienceFacet

A wrapper around the underlying SalienceFacet structure in the C API.

`sFacet`	The text of the facet
`nCount`	The count of occurrences of the facet
`nPositiveCount`	The count of positive occurrences of the facet, based on value of the Neutral Upper Bound option
`nNegativeCount`	The count of negative occurrences of the facet, based on value of the Neutral Lower Bound option
`nNeutralCount`	The count of neutral occurrences of the facet, based on value of the Neutral Lower Bound and Neutral Upper Bound options
`Attributes`	A List<> of SalienceAttribute structures for the facet
`nSubjectLength`	Number of SalienceAttribute structures for the facet
`Mentions`	A List<> of SaliencePhrase structures with details about the occurrences
`PositiveMentions`	A List<> of SaliencePhrase structures with details about the positive occurrences of the facet, based on the value of the Neutral Upper Bound option
`NegativeMentions`	A List<> of SaliencePhrase structures with details about the negative occurrences of the facet, based on the value of the Neutral Lower Bound option
`NeutralMentions`	A List<> of SaliencePhrase structures with details about the neutral occurrences of the facet, based on the value of the Neutral Upper Bound and Neutral Lower Bound options

SalienceIntention

A wrapper around the underlying SalienceIntention structure in the C API.

`sWhat`	The object of the intention, if detected. Otherwise, this structure member will be empty.
`sWho`	The expresser of the intention, if detected. Otherwise, this structure member will be empty.
`sEvidence`	The phrase expressing the intention.
`sType`	The intention type, out of the set of defined intention types, that was detected.
`oWhat`	A SalienceChunk containing positional information about the chunk identifying the object of the intention.
`oWho`	A SalienceChunk containing positional information about the chunk identifying the expresser of the intention. This will be an empty structure containing zeroes if acWho has not been detected.
`oEvidence`	A SalienceChunk containing positional information about the chunk containing the intention

SalienceMention

A wrapper around the underlying SalienceMention structure in the C API.

`Phrase`	A SaliencePhrase structure containing location of the mention in text
`nType`	The type of mention (Named, Nominal, Pronominal)
`fScore`	How likely it is that this is actually a mention of the entity

SalienceModelSentiment

A wrapper around the underlying SalienceSentimentModel structure in the C API.

`sName`	Name of the sentiment model used for scoring
`nBest`	Indicator of the sentiment category determined as best match (0=Positive, 1=Negative, 2=Mixed, 3=Neutral)
`fPositive`	Model-based score associated with likelihood document is positive
`fNegative`	Model-based score associated with likelihood document is negative
`fMixed`	Model-based score associated with likelihood document is mixed
`fNeutral`	Model-based score associated with likelihood document is neutral

Note: Sum of model scores should equal 1, nBest should match category of greatest model score.

SalienceOpinion

A wrapper around the underlying SalienceOpinion structure in the C API.

`Speaker`	A SalienceEntity structure for the entity expressing the opinion
`EntityTopic`	A SalienceEntity structure for the entity that is the subject of the opinion, if applicable
`ThemeTopic`	A SalienceTheme structure for the theme that is the subject of the opinion, if applicable
`sQuotation`	The text of the expressed opinion
`fSentiment`	The sentiment associated with the opinion
`bThemeOpinion`	An indicator of whether the opinion is being expressed about an entity or a theme

SaliencePhrase

A wrapper around the underlying SaliencePhrase structure in the C API.

`sText`	The text of the phrase
`nDocument`	The zero-based index of the document in the collection
`nSentence`	The zero-based index of the sentence
`nWord`	The zero-based index of the first token in the phrase
`nLength`	The length (in tokens) of the phrase
`nByte`	The byte offset in the original text that the phrase starts at
`nByteLength`	The length of the phrase in bytes
`nIsNegated`	An indicator of whether the phrase is negated (1) or not (0)
`sNegator`	If the phrase is negated, the negator is provided in this field
`nType`	An indicator of whether the phrase is negated (1), modified (2), both (3), or neither (0). This is only relevant to phrases related to SalienceSentimentPhrases.
`nSection`	The zero-based index of the document details section that this phrase occurs in
`nRow`	The zero-based index of the row if the phrase occurs in a list or table
`nColumn`	The zero-based index of the column if the phrase occurs in a list or table

SalienceRelationship

A wrapper around the underlying SalienceRelation structure in the C API.

`fScore`	Measure of confidence in the relationship
`sLabel`	Type of relationship, as defined in the relationship pattern
`sExtra`	Additional information associated with the relationship
`Entities`	A List<> of the SalienceEntity structures for the entities involved in the relationship

SalienceRow

A wrapper around the underlying SalienceRow structure in the C API.

`Columns`	A List<> of SalienceDocument structures representing the contents of the columns within the row of a table

SalienceSection

A wrapper around the underlying SalienceSection structure in the C API.

`sText`	The internal version of the document
`sFingerprint`	A calculated fingerprint of the document (DEPRECATED)
`sHeader`	The header for this section of the document, if applicable. Otherwise, blank.
`nWordCount`	The number of SalienceWords in the document
`nSentenceCount`	The number of SalienceSentences in the document
`nObjectiveCount`	The number of objective SalienceSentences in the document
`nSubjectiveCount`	The number of subjective SalienceSentences in the document
`nParsedCount`	The number of sentences that grammatically parse
`TermFrequency`	A List<> of SalienceToken structures providing term frequency analysis
`TaggedTermFrequency`	A List<> of SalienceToken structures providing tagged term frequency analysis
`BiGramFrequency`	A List<> of SalienceToken structures providing bi-gram frequency analysis
`TaggedBiGramFrequency`	A List<> of SalienceToken structures providing tagged bigram frequency analysis
`TriGramFrequency`	A List<> of SalienceToken structures providing tri-gram frequency analysis
`TaggedTriGramFrequency`	A List<> of SalienceToken structures providing tagged trigram frequency analysis
`QuadGramFrequency`	A List<> of SalienceToken structures providing quad-gram frequency analysis, for ideogram languages only
`oNegators`	A List<> of SalienceToken structures providing information about negators in the content
`oIntensifiers`	A List<> of SalienceToken structures providing information about intensifiers in the content
`Sentences`	A List<> of SalienceSentence structures for the sentences in the content
`Chunks`	A List<> of SalienceChunk structures providing chunk breakdown of the document
`Rows`	A List<> of SalienceRow structures for table and list-based content

SalienceSentence

A wrapper around the underlying SalienceSentence structure in the C API.

`Tokens`	A List<> of SalienceWord structures that make up the sentence
`Chunks`	A List<> of SalienceChunk structures that make up the sentence
`nLength`	The number of SalienceWords in the sentence
`nChunkCount`	The number of SalienceChunks in the sentence
`nSubjective`	An indicator specifying if the sentence is subjective (1) or objective (0)
`nPolar`	An indicator specifying if the sentence conveys sentiment polarity (1) or not (0)
`nImperative`	An indicator specifying if the sentence conveys an imperative (or "call to action") (1) or not (0)
`sText`	The text of the sentence
`nSummaryRank`	The rank of the sentence with relation to the summary. A call to retrieve the document summary must be made for this field to be initialized.
`fSentiment`	Reserved for future use to provide sentence-level sentiment. Currently returns 0.

SalienceSentiment

A wrapper around the underlying SalienceSentimentResult structure in the C API.

`fScore`	Phrase-based sentiment score
`Phrases`	A List<> of SalienceSentimentPhrase structures containing details of phrase-based sentiment analysis
`ModelSentiment`	A List<> of SalienceModelSentiment structures containing details of model-based sentiment analysis
`Emotions`	A List<> of SalienceTopics structures containing details of emotions detected in the content

SalienceSentimentPhrase

A wrapper around the underlying SalienceSentimentPhrase structure in the C API.

`Phrase`	A SaliencePhrase structure with details about the phrase
`sSource`	The path of the HSD file containing the phrase, if HSD-based
`fScore`	The sentiment weight for the phrase
`nType`	An indicator specifying if the phrase is HSD-based or not
`nModified`	An indicator of whether the phrase is negated (1), modified (2), both (3), or neither (0)
`SupportingPhrases`	A List<> of SaliencePhrase structures with details about any relevant supporting phrase, such as a negating or intensifying phrase

A SentimentPhrase has the following helpful functions for understanding the score:

getType() Returns an int, denoting how this phrase was identified:

0 : This was in the “lsf file”, which is a whole bunch of sentiment phrases already identified by Salience
1 : This was in an “hsd file”, either customer provided or general.hsd from Salience
Mostly that’s all you need, but the rest of the set is:
2 : This was defined in a special file “documentconcepts.dat”
3 : This was defined in a special file “metaphors.dat” for phrases longer than 5 words
4 : You turned “all possible sentiment phrases on” and while this isn’t a sentiment phrase, maybe it could be.
5 : You defined a query in your hsd file (“hot NOT WITH sun OR fire OR fever”)
6 : This phrase was found but isn’t being used because it’s in a question
7 : This phrase was found via a specially trained model.

getSource() For dictionary based results, returns the file name where this was found.

getModified() : Returns a bitmask for modifications to this sentiment phrase.

1 : The phrase was negated
2 : The phrase was intensified e.g. "very"
Those are the main ones, but also:
4 : The phrase was typographically intensified: currently just letter repetition e.g. "I looovvvveee this"
8 : The phrase was modified by its use in discourse e.g. “desert was good, but everything else was bad” the use of ‘but’ generally conveys greater emphasis for the latter idea over the former.
16 : The sentence ended in an exclamation mark

getSupportingPhrases() : this returns the Phrase objects which were the negators or intensifiers discovered.

SalienceStartup

A wrapper around the underlying SalienceSessionStartup structure in the C API.

`sLogPath`	Fully-qualified path to location where startup log can be written
`sDataDirectory`	Path to data directory for use
`sUserDirectory`	Path to user directory for use
`nMode`	A `SALIENCE_MODE` enum value indicating the mode for the Salience session

The nMode member may be set to the following values: SALIENCEMODE_DEFAULT: Default mode for handling general content. SALIENCEMODE_SHORTFORM: Mode for handling short-form content such as Twitter.

SalienceSummary

A wrapper around the underlying SalienceSummaryResult structure in the C API.

`sSummary`	Summary text generated for the document via default method
`vSentences`	A List<> of SalienceSentence objects providing the individual sentences in the summary with ranking
`sAlternateSummary`	Summary text generated for the document via an alternate method
`vAlternateSentences`	A List<> of SalienceSentence objects providing the individual sentences in the alternate summary with ranking

SalienceTheme

A wrapper around the underlying SalienceTheme structure in the C API.

`sTheme`	The text of the theme
`sStemmedTheme`	The stemmed version of the theme
`sNormalizedTheme`	The normalized version of the theme
`nThemeType`	An indicator is this is a "meta-theme" (1) or not (0)
`fScore`	Documents: A measure of the strength of the theme within the document. Collections: A count of the themes that have been rolled together across collection documents into this theme result
`fSentiment`	Documents: The sentiment score for the theme. Collections: A sentiment score derived from aggregating the sentences containing the theme mentions into a single document
`nEvidence`	Documents: A measure (from 1 to 7) of the content on which the sentiment score for the theme is based. Collections: The number of sentiment phrases used in calculating the theme's sentiment as described above
`bAbout`	Documents: An indicator specifying if the theme is contained within the summary of the document. Collections: False (not used)
`sSummary`	A summary of the document content relevant to the theme
`ThemeMentions`	A List<> of SaliencePhrase structures providing information about theme occurrences in the document
`Topics`	A List<> of SalienceTopic structures providing information about topics related to the theme
`AlternateThemes`	Reserved for future use.
`ChildMentions`	A List<> of SaliencePhrase structures containing information about child mentions of the theme.
`RelatedMentions`	A List<> of SaliencePhrase structures containing information about related mentions of the theme.

SalienceToken

A wrapper around the underlying SalienceToken structure in the C API.

`sToken`	A single token (generally a word, but also includes punctuation and other individual textual items)
`nCount`	Number of occurrences of the token

SalienceTopic

A wrapper around the underlying SalienceTopic structure in the C API. This structure is used to return results for multiple API calls:

Query-defined topics

`sTopic`	The query-defined topic label
`nHits`	Documents: This is the number of query terms that hit for the document. Collections: This is the number of documents within the collection that contain hits for the topic.
`fScore`	Documents: This is the sentiment score for content associated with topic query hits. Collections: This value is always 0 (not used).
`vDocuments`	Documents: A List<> of sentences containing the topic hits. Collections: A List<> of the document identifiers for the documents identified for the topic.
`nType`	0 (indicates query topic result)
`sSummary`	Documents: A summary of the document focused on the query result. Collections: Not used.
`vChildren`	Not used
`vEntities`	A List<> of entities that match the `{entity ...}` syntax if used in the query definition
`SentimentPhrases`	A List<> of SalienceSentimentPhrase structures of sentiment phrases related to this entity

Concept-defined topics

`sTopic`	The concept-defined topic label
`nHits`	Documents: This value is always 0 (not used). Collections: This is the number of documents within the collection that conceptually match for the topic.
`fScore`	Documents: This provides an indication of the strength of the match. Collections: This value is the sum of match strengths for the topic across all documents in the collection that match for the topic.
`vDocuments`	Documents: A List<> of sentences containing the topic hits. Collections: A List<> of the document identifiers for the documents identified for the topic.
`nType`	1 (indicates concept topic result)
`sSummary`	Documents: A summary of the document focused on the query result. Collections: Not used.
`vChildren`	Not used
`vEntities`	Not used
`SentimentPhrases`	A List<> of SalienceSentimentPhrase structures of sentiment phrases related to this entity

Document categories

Document categories are returned using the Wikipedia category information contained within the Concept Matrix. Document categories are only available at the document level.

`sTopic`	The concept-defined topic label
`nHits`	0 (not used)
`fScore`	This provides an indication of the strength of the match
`nType`	An integer indicating the type of category result: 2=category node, 3=category leaf, 4=category explain info
`vChildren`	A List<> of SalienceTopic objects representing child topics
`SentimentPhrases`	A List<> of SalienceSentimentPhrase structures of sentiment phrases related to this entity

Document classification

Document classification results are returned when using classification models in Salience 6. Document classification results are only available at the document level.

`sTopic`	The classification label
`fScore`	This provides an indication of the strength of the match for the classification

SalienceWord

A wrapper around the underlying SalienceWord structure in the C API.

`sToken`	The token associated with the word
`sPOSTag`	The part-of-speech tag associated with the token, see Supported POS tags
`sStem`	The stemmed version of the word
`bInvert`	An indicator specifying if sentiment for the word has been inverted
`fSentiment`	Sentiment for word, filled in when returned for sentiment and entity markup
`nSentimentType`	Indicator of source of sentiment; from hand-scored dictionary, or stopword
`sEntityType`	Entity type for word if it is part of an entity, returned for entity markup
`nID`	Identifiers of which markup applies to the word
`nSecondaryID`	Secondary identifier of which markup applies to the word
`bPostFixed`	An indicator specifying if the word is post-fixed (followed by a space)

Salience

The Salience Developer Hub

Data Structures