Save Classifier with Pickle - Natural Language Processing With Python and NLTK p.14
427As you will likely find with any form of data analysis, there is going to be some sort of processing bottleneck, that you repeat over and over, often yielding the same object in Python memory.
Examples of this might be loading a massive dataset into memory, some basic pre-processing of a static dataset, or, like in our case, the training of a classifier.
In our case, we spend much time on training our classifier, and soon we may add more. It is a wise choice to go ahead and pickle the trained classifer. This way, we can load in the trained classifier in a matter of milliseconds, rather than waiting 3-5+ minutes for the classifier to be trained.
To do this, we use the standard library's "pickle" module. What pickle does is serialize, or de-serialize, python objects. This could be lists, dictionaries, or even things like our trained classifier!
Playlist link: https://www.youtube.com/watch?v=FLZvOKSCkxY&list=PLQVvvaa0QuDf2JswnfiGkliBInZnIC4HL&index=1
sample code: http://pythonprogramming.net
http://hkinsley.com
https://twitter.com/sentdex
http://sentdex.com
http://seaofbtc.com