Sentiment

Sentiment Object

KEY	TYPE	DESCRIPTION
subsentences	list of Subsentences Sentiment Element Object	Source of all sentiments elements divided by subsentences
elements	list of Sentiment Element Object	Specific source for all sentiment elements in the sentence
values	Sentiment Values Object	normalized total values for sentiment

Sentiment values are available at the following levels of granularity: sentence, subsentence and element.

Subsentences Sentiment Object

KEY	TYPE	DESCRIPTION
sentence	string	Source of all sentiments elements divided by subsentences
start_id	int	Id of the first token of the subsentence
end_id	int	Id of the last token of the subsentence
elements	list of Sentiment Element Objets	Specific source for all sentiment elements in the subsentence
values	Sentiment Values Object	normalized values for sentiment of subsentence

Sentiment Element Object

KEY	SUBKEY	TYPE	DESCRIPTION	CONSTRAINTS
source	index	int	Index of the source word for this element	-
-	lemma	string	source's lemma	-
-	lemma	string	source's source	-
subject	index	int	Index of the subject for this element	-
-	lemma	string	subject's lemma	-
-	source	string	subject's source	-
target	index	int	Index of the target for this element	-
-	lemma	string	target's lemma	-
-	source	string	target's source	-
value	-	float	Indice of the sentiment	-1 < value < 1

Subject is the word which represents the subject of the action1

Source is the word which conveys the sentiment

Target is the word which represents the target of the sentiment

Sentiment Values Object

KEY	TYPE	DESCRIPTION	CONSTRAINTS
positive	float	normalized addition of all positive sentiment values in the sentence	0 <= positive < 1
negative	float	normalized addition of all negative sentiment values in the sentence	-1 < negative <= 0
total	float	positive + negative	-1 < total < 1

Values are calculated either by using sentiment elements objects if available, or by a prediction model at the subsentence level. Values are normalized to stay in the the [-1 : 1] interval between element, subsentence and sentence level therefore comparisons should be made made with elements of the same depth.

import Tabs from '@theme/Tabs'; import TabItem from '@theme/TabItem';

{
    "elements": [
        {
            "source": {
                "index": 1,
                "lemma": "aimer",
                "source": "aime"
            },
            "subject": {
                "index": 0,
                "lemma": "il",
                "source": "elle"
            },
            "target": {
                "index": 3,
                "lemma": "chien",
                "source": "chiens"
            },
            "value": 0.46
        }
    ],
    "values": {
        "negative": 0,
        "positive": 0.17,
        "total": 0.17
    },
    "subsentences": {
        "sentence": "elle aime les chiens",
        "start_id": 0,
        "end_id": 3,
        "elements": [
            {
                "source": {
                    "index": 1,
                    "lemma": "aimer",
                    "source": "aime"
                },
                "subject": {
                    "index": 0,
                    "lemma": "il",
                    "source": "elle"
                },
                "target": {
                    "index": 3,
                    "lemma": "chien",
                    "source": "chiens"
                },
                "value": 0.27
            }
        ],
        "values": {
                "negative": 0,
                "positive": 0.17,
                "total": 0.17
        }
    }
}

</TabItem> <TabItem value="json" label="JSON">

{
	"elements": [
		{
			"source": {
				"index": 1,
				"lemma": "aimer",
				"source": "aime"
			},
			"subject": {
				"index": 0,
				"lemma": "il",
				"source": "elle"
			},
			"target": {
				"index": 3,
				"lemma": "chien",
				"source": "chiens"
			},
			"value": 0.46
		}
	],
	"values": {
		"negative": 0,
		"positive": 0.17,
		"total": 0.17
	},
	"subsentences": {
		"sentence": "elle aime les chiens",
		"start_id": 0,
		"end_id": 3,
		"elements": [
			{
				"source": {
					"index": 1,
					"lemma": "aimer",
					"source": "aime"
				},
				"subject": {
					"index": 0,
					"lemma": "il",
					"source": "elle"
				},
				"target": {
					"index": 3,
					"lemma": "chien",
					"source": "chiens"
				},
				"value": 0.27
			}
		],
		"values": {
			"negative": 0,
			"positive": 0.17,
			"total": 0.17
		}
	}
}

</TabItem> </Tabs>