Bag of Tricks for Efficient Text Classification
This provides a highly efficient solution for large-scale text classification tasks, though it is incremental as it builds on existing bag-of-words and linear classifier methods.
The paper tackled efficient text classification by introducing fastText, a simple baseline that matches deep learning classifiers in accuracy while being orders of magnitude faster, training on over one billion words in under ten minutes and classifying half a million sentences among 312K classes in under a minute.
This paper explores a simple and efficient baseline for text classification. Our experiments show that our fast text classifier fastText is often on par with deep learning classifiers in terms of accuracy, and many orders of magnitude faster for training and evaluation. We can train fastText on more than one billion words in less than ten minutes using a standard multicore~CPU, and classify half a million sentences among~312K classes in less than a minute.