Leveraging text profiles to select and configure models for use with textual datasets

A method comprising:

• analyzing, by one or more processors, a reference textual dataset by applying a plurality of text-mining techniques to the reference textual dataset to generate a first text profile containing a first plurality of metrics characterizing the reference textual dataset;
• analyzing, by the one or more processors, a target textual dataset by applying the plurality of text-mining techniques to generate a second text profile containing a second plurality of metrics characterizing the target textual dataset;
• determining, by the one or more processors, one or more similarity values representing how similar the target textual dataset is to the reference textual dataset by comparing the second text profile to the first text profile, wherein comparing the second text profile to the first text profile involves comparing at least some of the second plurality of metrics to at least some of the first plurality of metrics;
• receiving, from a user and by the one or more processors, a selection of a model that is to be applied to the target textual dataset;
• in response to receiving the selection, generating, by the one or more processors, one or more insights relating to an anticipated accuracy of the model on the target textual dataset based on the one or more similarity values and one or more characteristics of the model;
• and
• outputting, by the one or more processors, the one or more insights to the user.

A non-transitory computer-readable medium comprising program code that is executable by one or more processors for causing the one or more processors to:

• analyze a reference textual dataset by applying a plurality of text-mining techniques to the reference textual dataset to generate a first text profile containing a first plurality of metrics characterizing the reference textual dataset;
• analyze a target textual dataset by applying the plurality of text-mining techniques to generate a second text profile containing a second plurality of metrics characterizing the target textual dataset;
• determine one or more similarity values representing how similar the target textual dataset is to the reference textual dataset by comparing the second text profile to the first text profile, wherein comparing the second text profile to the first text profile involves comparing at least some of the second plurality of metrics to at least some of the first plurality of metrics;
• receive, from a user, a selection of a model that is to be applied to the target textual dataset;
• in response to receiving the selection, generate one or more insights relating to an anticipated accuracy of the model on the target textual dataset based on the one or more similarity values and one or more characteristics of the model;
• and
• output the one or more insights to the user.