Skip to content

What to train the classifier on? #6

@omstuhler

Description

@omstuhler

In the previous week, we used different classification methods to assign documents to categories. We trained our classifiers on the word vectors of the documents in our training dataset. In the section on relation extraction via supervised learning, Jurafsky and Martin point to a number of different features beyond mere word vectors that can be used for this type of classification task such as word positions or dependency tree paths. I wonder how exactly we simultaneously incorporate these features into our data structure or, put in practical terms: how do the vectors look like that we feed a classifier with for supervised relation extraction.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions