Text classifiers can learn patterns that we don’t intend for them to learn.We’ve talked about evaluating a classifier on held-out data using standard metrics like precision, recall, and f-score. These metrics are…Mar 14, 2021Mar 14, 2021
Different classifiers use different algorithms to learn how to predict document categories.So far we’ve seen the basic process for building a text classifier:Mar 7, 2021Mar 7, 2021
Sentiment analysis provides a final piece of meaning: the opinion expressed in a text.(Copy Paste)Feb 7, 2021Feb 7, 2021
The structural parts of language contain information about authorship and style.(collect)Feb 2, 2021Feb 2, 2021
We can start to represent the meaning of a text by counting words that carry content.(collect)Jan 31, 2021Jan 31, 2021
We measure how well a classifier works by comparing predicted labels against actual labels.(Copy Paste)Jan 11, 2021Jan 11, 2021
New technology needs to be understood using existing ethical principles.Artificial intelligence allows us to access knowledge that we as individuals could never hope to analyze and read ourselves. But this…Jan 9, 2021Jan 9, 2021
Pipelines allow us to automate our experiments from data collection all the way to visualizationLet’s say we want to monitor corruption in societies and governments. The more corruption there is, we think, the more people will be…Jan 6, 2021Jan 6, 2021
The problem of text classificationFirst, we need to design a category system. If I want to sort hotels by the kind of guests they cater to, I need to make up a set of hotel…Jan 2, 2021Jan 2, 2021
