Doc2VecVectorizer<T>

Models & Types ยท AiDotNet.Preprocessing.TextVectorizers

Converts text documents to dense vectors using Doc2Vec (Paragraph Vector).

Inheritance: TextVectorizerBase<T> โ†’ Doc2VecVectorizer<T>

For Beginners

Doc2Vec is Word2Vec for whole documents:

  • Each document gets its own unique learned vector
  • Captures document meaning, not just word averages
  • Great for document similarity, classification, and clustering
  • Better than averaging word vectors for longer documents

How It Works

Doc2Vec extends Word2Vec to learn fixed-length representations for documents. Two architectures are supported:

  • PV-DM (Distributed Memory): Uses document vector with context words to predict target
  • PV-DBOW (Distributed Bag of Words): Uses document vector alone to predict words

Unlike Word2Vec averaging, Doc2Vec learns document-specific vectors that capture document-level semantics beyond just word averages.

Constructors

Constructor Summary
Doc2VecVectorizer(int vectorSize, int windowSize, int minCount, int epochs, double learningRate, int negativeSamples, Doc2VecArchitecture architecture, bool lowercase, int? randomState, Func<string, IEnumerable<string>> tokenizer, HashSet<string> stopWords, ITokenizer advancedTokenizer) Creates a new instance of Doc2VecVectorizer.

Properties

Property Summary
Dictionary<int, double[]> DocVectors { get; } Gets the learned document vectors (indexed by training order).
Dictionary<string, double[]> WordVectors { get; } Gets the learned word vectors.
bool IsFitted { get; }
int FeatureCount { get; }
int VectorSize { get; } Gets the vector dimensionality.

Methods

Method Summary
Matrix<T> Transform(IEnumerable<string> documents) Transforms documents to dense vectors by inferring document vectors.
void Fit(IEnumerable<string> documents) Trains Doc2Vec embeddings on the corpus.