Pairwise similarity of TopSig document signatures

Vries, Christopher M. De; Geva, Shlomo

doi:10.1145/2407085.2407103

Cited by 9 publications

(10 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Early approaches used to generate document signatures are based on the bitwise OR composition of binary signatures associated with terms in documents [9,7]. Further refinements of the signature generation process have been proposed.…”

Section: Patent Classification With Signaturesmentioning

confidence: 99%

“…Further refinements of the signature generation process have been proposed. A recent approach, called TopSig [10,7], uses random indexing for compressing the standard term-document matrix, followed by aggressive numerical precision reduction to maintain only the sign bits of the projected term-document matrix. This approach has been shown to be superior to standard signature approaches: we thus rely on the TopSig method to generate patent signatures.…”

Section: Patent Classification With Signaturesmentioning

confidence: 99%

See 1 more Smart Citation

A Signature Approach to Patent Classification

Seneviratne

Geva

Zuccon

et al. 2015

Information Retrieval Technology

Self Cite

View full text Add to dashboard Cite

Abstract. We propose a document signature approach to patent classification. Automatic patent classification is a challenging task because of the fast growing number of patent applications filed every year and the complexity, size and nested hierarchical structure of patent taxonomies. In our proposal, the classification of a target patent is achieved through a k-nearest neighbour search using Hamming distance on signatures generated from patents; the classification labels of the retrieved patents are weighted and combined to produce a patent classification code for the target patent. The use of this method is motivated by the fact that, intuitively, document signatures are more efficient than previous approaches for this task that considered the training of classifiers on the whole vocabulary feature set. Our empirical experiments also demonstrate that the combination of document signatures and k-nearest neighbours search improves classification effectiveness, provided that enough data is used to generate signatures.

show abstract

Section: Patent Classification With Signaturesmentioning

confidence: 99%

Section: Patent Classification With Signaturesmentioning

confidence: 99%

A Signature Approach to Patent Classification

Seneviratne

Geva

Zuccon

et al. 2015

Information Retrieval Technology

Self Cite

View full text Add to dashboard Cite

show abstract

“…In order to employ RMI for the construction of a VSM at reduced dimension, two model parameters should be decided: (a) the dimension of the VSM, which is shown by m, and (b) the number of non-zero elements in index vectors, which is determined by s in Equation 5. In contrast to the classic one-dimension-per-context-element methods of VSM construction, 5 the value of m in RPs and thus in RMI is chosen independently of the number of context elements n in the model.…”

Section: Random Manhattan Indexingmentioning

confidence: 99%

“…In contrast to the classic one-dimension-per-context-element methods of VSM construction, 5 the value of m in RPs and thus in RMI is chosen independently of the number of context elements n in the model. In RMI, as shown in [18], m is established by the probability and the maximum expected amount of distortions in pairwise distances and the number of vectors p in the model: a larger m yields to lower bounds on the distortion with a higher probability, i.e.…”

Section: Random Manhattan Indexingmentioning

confidence: 99%

See 1 more Smart Citation

Random Manhattan Indexing

Zadeh

Handschuh

2014

2014 25th International Workshop on Database and Expert Systems Applications

View full text Add to dashboard Cite

Abstract-Vector space models (VSMs) are mathematically well-defined frameworks that have been widely used in text processing. In these models, high-dimensional, often sparse vectors represent text units. In an application, the similarity of vectorsand hence the text units that they represent-is computed by a distance formula. The high dimensionality of vectors, however, is a barrier to the performance of methods that employ VSMs. Consequently, a dimensionality reduction technique is employed to alleviate this problem. This paper introduces a new method, called Random Manhattan Indexing (RMI), for the construction of 1 normed VSMs at reduced dimensionality. RMI combines the construction of a VSM and dimension reduction into an incremental, and thus scalable, procedure. In order to attain its goal, RMI employs the sparse Cauchy random projections.

show abstract

Random Indexing Revisited

QasemiZadeh

2015

Natural Language Processing and Information Systems

View full text Add to dashboard Cite

Random indexing is a method for constructing vector spaces at a reduced dimensionality. Previously, the method has been proposed using Kanerva's sparse distributed memory model. Although intuitively plausible, this description fails to provide mathematical justification for setting the method's parameters. The random indexing method is revisited using the principles of sparse random projections in Euclidean spaces in order to complement its previous delineation.

show abstract

Pairwise similarity of TopSig document signatures

Cited by 9 publications

References 8 publications

A Signature Approach to Patent Classification

A Signature Approach to Patent Classification

Random Manhattan Indexing

Random Indexing Revisited

Contact Info

Product

Resources

About