Improving Text Classification Performance Using PCA and Recall-Precision Criteria

Zahedi, Morteza; Sorkhi, Ali Ghanbari

doi:10.1007/s13369-013-0569-2

Cited by 14 publications

(6 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Apart from mathematics, researchers also utilize the properties of eigenvalue and eigenvector in Schrdinger equation [6], geology [11] and vibration analysis. The widely known application would be Principal Component Analysis(PCA) [13], which is used in dimension reduction [14], feature selection [20,24,12], K-means clustering [7] and general text classification problems [10,28,26,22].…”

Section: Related Workmentioning

confidence: 99%

An Analytical Formula for Spectrum Reconstruction

Dai,

Matzinger,

Popescu

2020

Preprint

View full text Add to dashboard Cite

We study the spectrum reconstruction technique. As is known to all, eigenvalues play an important role in many research fields and are foundation to many practical techniques such like PCA(Principal Component Analysis). We believe that related algorithms should perform better with more accurate spectrum estimation. There was an approximation formula proposed by[1], however, they didn't give any proof. In our research, we show why the formula works. And when both number of features and dimension of space go to ∞, we find the order of error for the approximation formula, which is related to a constant c-the ratio of dimension of space and number of features.

show abstract

Section: Related Workmentioning

confidence: 99%

An Analytical Formula for Spectrum Reconstruction

Dai,

Matzinger,

Popescu

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…If the text is longer or if it has ambiguous semantic feature items, then the frequency weight (Salton and Buckley, 1988) could be used. Further, when the distribution of feature items in texts is considered, TF-IDF (Chen et al , 2016; Zahedi and Sorkhi, 2013) will be adopted to avoid the reproducibility of feature items in texts. Additional TF factors for describing the relationships between categories and features include inverse class frequency (ICF) (Peng et al , 2014; Ren and Sohrab, 2013) and MI, among other methods.…”

Section: Related Workmentioning

confidence: 99%

A paper-text perspective

Wang

Deng

2017

View full text Add to dashboard Cite

Purpose In the era of Big Data, network digital resources are growing rapidly, especially the short-text resources, such as tweets, comments, messages and so on, are showing a vigorous vitality. This study aims to compare the categories discriminative capacity (CDC) of Chinese language fragments with different granularities and to explore and verify feasibility, rationality and effectiveness of the low-granularity feature, such as Chinese characters in Chinese short-text classification (CSTC). Design/methodology/approach This study takes discipline classification of journal articles from CSSCI as a simulation environment. On the basis of sorting out the distribution rules of classification features with various granularities, including keywords, terms and characters, the classification effects accessed by the SVM algorithm are comprehensively compared and evaluated from three angles of using the same experiment samples, testing before and after feature optimization, and introducing external data. Findings The granularity of a classification feature has an important impact on CSTC. In general, the larger the granularity is, the better the classification result is, and vice versa. However, a low-granularity feature is also feasible, and its CDC could be improved by reasonable weight setting, even exceeding a high-granularity feature if synthetically considering classification precision, computational complexity and text coverage. Originality/value This is the first study to propose that Chinese characters are more suitable as descriptive features in CSTC than terms and keywords and to demonstrate that CDC of Chinese character features could be strengthened by mixing frequency and position as weight.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Each method has its applications in textual processing as in Zahedi et al (2012) and Underhill et al (2007). In Zahedi et al (2012), the paper uses PCA and Precision/Recall criteria for text processing. Underhill et al (2007) compares 5 dimension reduction methods to discuss their advantages and disadvantages in textual processing.…”

Section: Related Workmentioning

confidence: 99%

“…We can see that PCA can identify the principal components and transform the original scores into the principal components' coordinate system to strengthen the weight of the principal components. The text classification is improved using PCA and Recall-Precision Criteria in Zahedi et al (2012). In another hand, the result of the FA is to summarize the influence of the main factors on the documents to be reflected in the FA loadings.…”

Section: Hybrid Matrix Factorization and Optimizationmentioning

confidence: 99%

See 1 more Smart Citation