{"id":604,"date":"2021-04-15T19:40:37","date_gmt":"2021-04-15T23:40:37","guid":{"rendered":"https:\/\/www.macloo.com\/ai\/?p=604"},"modified":"2021-04-15T19:40:37","modified_gmt":"2021-04-15T23:40:37","slug":"loving-the-spacy-tutorial-for-nlp","status":"publish","type":"post","link":"https:\/\/www.macloo.com\/ai\/2021\/04\/15\/loving-the-spacy-tutorial-for-nlp\/","title":{"rendered":"Loving the spaCy tutorial for NLP"},"content":{"rendered":"\n

I got my first look at spaCy<\/a>, a Python library for natural language processing, near the end of 2019. I wanted to learn it but had too many other things to do. Fast-forward to now, almost 14 months into the pandemic, and I recently stumbled across spaCy’s own tutorial<\/a> for learning to use the library.<\/p>\n\n\n\n

The interactive tutorial includes videos, slides, and code exercises, and there is a GitHub repo<\/a>. It is available in English, Deutsch, Espa\u00f1ol, Fran\u00e7ais, Portugu\u00eas, \u65e5\u672c\u8a9e, and \u4e2d\u6587. Today I completed chapter 2. If you already know Python at, say, an intermediate level,<\/em> check it out!<\/p>\n\n\n\n

\"Screenshot
Trying out spaCy’s displaCy module and named entities.<\/em> <\/figcaption><\/figure><\/div>\n\n\n\n

In chapter 1 (there are four chapters), I got a handle on part-of-speech tags, syntactic dependencies, and named entities. I learned that we can search on these, and also on words (tokens<\/strong>) related to combinations<\/em> that we define. I’ve known about large-scale document searches (where a huge collection of documents is searched programmatically, usually to extract the most meaningful docs for some purpose \u2014 like a journalism investigation), and now I was getting a much better idea of how such searches can be designed.<\/p>\n\n\n\n

SpaCy provides “pre-trained model<\/strong> packages,” meaning someone else has already done the hard work of machine learning\/training to generate word vectors.<\/strong> There are packages of various sizes and in various languages<\/a>. Loading a model provides various features (the bigger the model, the more features). <\/p>\n\n\n\n

I think I was hooked as soon as I saw this and realized you could ask for all the MONEY entities, or all the ORG entities, in a document and evaluate them:<\/p>\n\n\n\n

\"\"
An example from chapter 1 in the spaCy tutorial.<\/em><\/figcaption><\/figure><\/div>\n\n\n\n

Then (still in chapter 1) I learned that I can easily define my own entities<\/em> if the model doesn’t recognize the ones I need to find. I learned that if I don’t know what GPE is, I can enter spacy.explain(\"GPE\")<\/code> and spaCy will return 'Countries, cities, states'<\/code> \u2014 sweet! <\/p>\n\n\n\n

Then I learned about rule-based matching,<\/strong> and I thought: “Regular expressions, buh-bye!”<\/p>\n\n\n\n

Chapter 1 didn’t really get deeply into lemmatization,<\/strong> but it offered this: <\/p>\n\n\n\n

\"\"
Lemmatization groups all forms of a word together so they can be analyzed as one item.<\/em><\/figcaption><\/figure><\/div>\n\n\n\n

That was just chapter 1! Chapter 2 went further into creating your own named entities and using parts of speech as part of your search criteria. For example, if you want to find all instances where a particular entity (say, a city) is followed by a verb \u2014 any verb \u2014 you can do that. Or any part of speech. You can construct a complex pattern, mixing specific words, parts of speech, and selected types of entities. The pattern can include as many tokens as you want. (If you’re familiar with regex<\/strong> \u2014 all the regex things are available.)<\/p>\n\n\n\n

You can determine whether phrases or sentences are similar<\/strong> to each other (although imperfectly).<\/p>\n\n\n\n

I’m not entirely sure how I would use these, but I’m sure they’re good for something:<\/p>\n\n\n\n