Using LSI for text classification in the presence of background text

Zelikovitz, Sarah; Hirsh, Haym

doi:10.1145/502585.502605

Cited by 107 publications

(41 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For data set 2, we compared the performance by varying the number of SVD dimensions k from 10,20,40,60,80,100,120,150,200,250,300,350,400,450, 500, 550 to 600. The neural network's input nodes number is equal to the dimensions of the document vectors.…”

Section: Experimental Results and Analysismentioning

confidence: 99%

“…LSA has been applied to text categorization in many previous works. Yang [19] used SVD for noise reduction to improve the computational efficiency in text categorization, Zelikovitz and Hirsh [20] performed LSA expanded term by document matrix in conjunction with background knowledge in text categorization, more recently, the supervised LSA [21] has been proposed to improve the performance in text categorization, and the idea of using latent semantic indexing and recurrent neural network for text classification has been mentioned by Mitra, et al [22].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Text categorization based on combination of modified back propagation neural network and latent semantic analysis

Wang

2008

Neural Comput & Applic

View full text Add to dashboard Cite

This paper proposed a new text categorization model based on the combination of modified back propagation neural network (MBPNN) and latent semantic analysis (LSA). The traditional back propagation neural network (BPNN) has slow training speed and is easy to trap into a local minimum, and it will lead to a poor performance and efficiency. In this paper, we propose the MBPNN to accelerate the training speed of BPNN and improve the categorization accuracy. LSA can overcome the problems caused by using statistically derived conceptual indices instead of individual words. It constructs a conceptual vector space in which each term or document is represented as a vector in the space. It not only greatly reduces the dimension but also discovers the important associative relationship between terms. We test our categorization model on 20-newsgroup corpus and reuter-21578 corpus, experimental results show that the MBPNN is much faster than the traditional BPNN. It also enhances the performance of the traditional BPNN. And the application of LSA for our system can lead to dramatic dimensionality reduction while achieving good classification results.

show abstract

Section: Experimental Results and Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Text categorization based on combination of modified back propagation neural network and latent semantic analysis

Wang

2008

Neural Comput & Applic

View full text Add to dashboard Cite

show abstract

“…Unlabeled examples have been used in conjunction with positive examples to improve classification (Li and Liu, 2003;Yu et al, 2003). Empirically it has been shown that combining labeled and unlabeled examples can improve accuracy on unseen test sets in certain situations (Blum and Mitchell, 1998;Mitchell, 1999;Joachims, 1999;Jaakkola et al, 1999;Zelikovitz and Hirsh, 2001;Bruce, 2001;Zhang and Oles, 2000).…”

Section: Related Workmentioning

confidence: 99%

Extending WHIRL with background knowledge for improved text classification

2006

Self Cite

View full text Add to dashboard Cite

Intelligent use of the many diverse forms of data available on the Internet requires new tools for managing and manipulating heterogeneous forms of information. This paper uses WHIRL, an extension of relational databases that can manipulate textual data using statistical similarity measures developed by the information retrieval community. We show that although WHIRL is designed for more general similarity-based reasoning tasks, it is competitive with mature systems designed explicitly for inductive classification. In particular, WHIRL is well suited for combining different sources of knowledge in the classification process. We show on a diverse set of tasks that the use of appropriate sets of unlabeled background knowledge often decreases error rates, particularly if the number of examples or the size of the strings in the training set is small. This is especially useful when labeling text is a labor-intensive job and when there is a large amount of information available about a particular problem on the World Wide Web.

show abstract

“…Principal components analysis is a simple method and has been proved to be useful to reduce features (Lam and Lee 1999). Latent semantic indexing is an amazing approach both for feature selection and reduction (Zelikovitz andHirsh 2001, Song and). More recently, natural language processing techniques, e.g.…”

Section: Introductionmentioning

confidence: 99%

Intelligent information retrieval system using automatic thesaurus construction

Song

et al. 2011

International Journal of General Systems

View full text Add to dashboard Cite

This paper presents an intelligent information retrieval (IR) system based on automatic thesaurus construction for its applications of document clustering and classification. These two applications are the most influential and widely used fields amongst the IR research community. We apply two biologically inspired algorithms, i.e. genetic algorithm (GA) and neural network (NN), to these two fields. A fuzzy logic controller GA and an adaptive back-propagation NN are proposed in our study, which can validly overcome the problems existing in their archetypes, e.g. slow evolution and being prone to trap into a local optimum. Furthermore, a well-constructed thesaurus has been recognised as a valuable tool in the effective operation of clustering and classification. It solves the problem in document representation organised by a bag of words, where some important relationships between words, e.g. synonymy and polysemy, are ignored. To investigate how our IR system could be used effectively, we conduct experiments on four data sets from the benchmark Reuter-21578 document collection and 20-newsgroup corpus. The results reveal that our IR system enhances the performance in comparison with k-means, common GA, and conventional backpropagation NN.

show abstract

Using LSI for text classification in the presence of background text

Cited by 107 publications

References 9 publications

Text categorization based on combination of modified back propagation neural network and latent semantic analysis

Text categorization based on combination of modified back propagation neural network and latent semantic analysis

Extending WHIRL with background knowledge for improved text classification

Intelligent information retrieval system using automatic thesaurus construction

Contact Info

Product

Resources

About