Intent Classification

The intent classifier is a specialized text classifier that analyses media posts about future actions. It will detect whether the post intends to perform some behavior, such as purchasing or selling and can be used alongside other classifiers and filters.

Statistics

Example Use Cases

The Intent classifier is designed to uncover instances where people express their intentions to perform actions, especially actions in the future. It is meant to quickly add additional information to help users assess the imminence of activities and identify future actions and interests relating to products or people (named entities). It could also detect whether future actions may be violent (when used in conjunction with the violence classifier).

Streaming Usage

Compatible Data Scoures

As a stream-integrated classifier, it is run on ingestion for specific sources.

📘

Recipe Available

View the below recipe to see it in action, and easily view how to integrate it into your own data pipeline.

Output

The Intent classifier is a binary label. The label value is ‘yes’ if the intent is detected in a post, ‘no’ otherwise. Included is a confidence score of 0.00 to 1.00.

"enrichment": 
            { 
						"intent": {
                    "label": "yes",
                    "confidence": 0.87
                },
}

Post-Processing Usage

Compatible Data Sources

As a Post-Processing operation, it can be run on any data source.

📘

Recipe Available

View the below recipe to see it in action, and easily view how to integrate it into your own data pipeline.

Usage

This Operation allows a user to specify the destination field, source fields, and any separator.

{
    "query": {
		...
},
    "operations": [
        {
            "name": "intent",
            "destination_path": "operations.intent",
            "parameters": {
                "language": "enrichment.language",
                "main": "content.body"
            }
        }
    ]
}
```