Latent semantic analysis
Latent semantic analysis is a dimensionality-reduction technique that factorizes term-document statistics to represent words and documents in a lower-dimensional semantic space.
Natural Language Processing in Action, Second Edition
Latent semantic indexing, also called latent semantic analysis, is a dimensionality-reduction technique that represents words and documents in a lower-dimensional semantic space derived from term-document statistics.
- How it is used
- Turns term statistics into topic vectors for semantic search, document comparison, and classification.
- How deep it goes
- Explains dimensionality reduction through co-occurrence and a TF–IDF classification example, connecting the resulting vectors to PCA and SVD.
- Examples
- Uses 5,000 Wikipedia discussion comments, a 19,169-feature TF–IDF matrix, and a comparison of topic-based toxicity classifiers.
- What you can do after
- Apply it via scikit-learn TruncatedSVD to produce topic features and evaluate their usefulness for document classification.
Read how Natural Language Processing in Action, Second Edition explains latent semantic analysis →
Math and Architectures of Deep Learning
A document-retrieval technique that uses low-dimensional structure in document-term data to capture latent topics and relationships.
- How it is used
- Uses dimensionality reduction and topic projection to improve document retrieval beyond direct term overlap.
- How deep it goes
- Explains the PCA and SVD basis, geometric compression, rank truncation, and projected cosine similarity through a toy computation.
- Examples
- Uses a four-term toy matrix and compares two documents before and after projection into one topic dimension.
- What you can do after
- Apply it via PyTorch to decompose a document-term matrix, reduce its rank, project documents, and compare topic-space vectors.
Read how Math and Architectures of Deep Learning explains latent semantic analysis →
Related concepts
- matrix factorizationuses · Matrix factorization is the decomposition or approximation of a matrix as a product of simpler, often lower-dimensional or structured matrices.
- term frequency-inverse document frequencyuses · Term frequency-inverse document frequency is a text-weighting scheme that combines a term’s frequency in a document with its inverse frequency across a document collection.
- singular value decompositionuses · Singular value decomposition is a matrix factorization that expresses a matrix as the product of orthogonal or unitary matrices and a diagonal matrix of nonnegative singular values.
- dimensionality reductionkind of · Dimensionality reduction is the transformation of data into a representation with fewer dimensions while retaining selected information, structure, or relationships.
- principal component analysisused by · Principal component analysis is a linear dimensionality-reduction technique that transforms data into orthogonal components ordered by the variance they capture.
- topic vectorproduces · A numerical vector encoding the strengths of associations between terms or documents and latent topics.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.