Tell us about bag-of-words and its limitations.

Bag-of-Words (BoW) is a simple method of representing text as a set of words without considering the order and grammatical structure. The text is broken down into words (tokens), and a vector is created where each element reflects the count or presence of a specific word in the document.

Limitations of BoW:

• Loss of word order. Context and meaning dependent on word order are not considered.
• Ignoring semantics. Words with different meanings but similar forms are treated the same.
• High dimensionality. For large vocabularies, vectors become sparse and bulky.
• Sensitivity to noise. Frequently occurring but uninformative words (stop-words) can influence the model.

Example of BoW for the sentence "Кошка сидит на ковре":

word	count
Кошка	1
сидит	1
на	1
ковре	1