Complex-Input Classification Task
(Redirected from labeling)
A Complex-Input Classification Task is a complex classification task that requires the classifying of complex objects.
- AKA: Tagging, Labeling, Structured Data Classification.
- a Data Object (with interrelated members).
- a Finite Set.
- a Classification Function, possibly generated by a Complex Input Supervised Classification System.
- output: a Class Member for each Data Item in the Data Object.
- It can be solved by a Complex-Input Classification System (that implements a Structured Input Classification Algorithm/Tagging Algorithm.
- It can range from being a Heuristic Tagging Task to being a Data-Driven Tagging Task (such as supervised tagging).
- It can range from being a Supervised Complex-Input Classification Task to being an Unsupervised Complex-Input Classification Task.
- It can range from being an Object Classification Task, to being a Component Classification Task, to being a Relationship Classification Task.
- It can be supported by a Recognition Task.
- It can require the Detection and Classification of Structure that is only implicitly present in an Artifact.
- It can be instantiated in a Complex-Input Classification Act.
- an Image Tagging Task.
- a Text Tagging Task.
- a Sequence Classification Task, such as a text classification task.
- a Sequence Member Classification Task, such as a DNA string tagging task or a part-of-speech tagging task.
- a Graph Classification Task.
- a Song Tagging Task.
- See: Recognition Task, Non-IID Relation.