![]() A similarity measure between real valued vectors (like cosine or euclidean distance) can thus be used to measure how words are semantically related. Explaining in more details how it is done is out of scope of this article, but the key points are that a precomputed model gives us real valued vectors associated to words and that those vectors contain a lot of information because the model is trained on a big dataset. The meaning of a word is learned from its surrounding words in the sentences and encoded in a vector of real values. Word2vec is a shallow neural network trained on a large text corpus. Word2vec, introduced in Distributed Representations of Words and Phrases and their Compositionality (Mikolov et al., NIPS 2013), has attracted a lot of attention in recent years due to its efficiency to produce relevant word embeddings (i.e. Computing semantic relationships between textual data enables to recommend articles or products related to a given query, to follow trends, to explore a specific subject in more details, etc.But texts can be very different miscellaneous: a Wikipedia article is long and well written, tweets are short and often not grammatically correct. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |