Understanding natural language processing
Text classification is an NLP task that analyzes and categorizes text into groups using predefined labels. Common business use cases for text classification include sentiment analysis, topic detection, and language detection. Classifying content provides valuable business insights into customer preferences, personalized experience, content moderation, emerging market segments, and social sentiment.
For an overview of NLP, we will train a multiclass text classification model using AutoGluon, https://auto.gluon.ai/, and a public data repository for Amazon customer reviews, https://doi.org/10.7910/DVN/W96OFO, maintained by Harvard Dataverse. The data repository contains 7 text datasets collected between 2008 and 2020, with 5,000 reviews each. You can download the export_food.csv
file from this data repository for the example walk through.
The following section gives an overview of AutoGluon.
Reviewing AutoGluon
AutoGluon is an open...