3 ms·
Natural Language Processing with Python updated for Python 3 and NLTK 3
- amirouche 9y agoI don't find nltk book to be a good introduction to solving NLP problems of today, it's good to introduce the low level tasks (like tokenizing, POS tagging, etc...) but nothing about summarization, Q/A etc.
- stared 9y agoUnless something changed (a lot) with NLTK, I had much better experience using spaCy https://spacy.io/ https://spacy.io/ (API and documentation-wise; and quality of results-wise).
- Xeoncross 9y agoI really like spaCy's design of throwing out old research and sticking with the best current methods, rather than leaving a dozen options to confuse newcomers. Having options is nice, but at least mark/default to the recommended ones.
- brian_spiering 9y agoThis is not a good resource for NLP. It is extremely dated (very little Deep Learning). I maintain a more update to list of resources https://github.com/brianspiering/awesome-dl4nlp https://github.com/brianspiering/awesome-dl4nlp Additionally, the nltk has an awkward API. TextBlob has similarity functionality with more intuitive OOP API.