INCREASING PERFORMANCE OF TEXT CLASSIFICATION BASED ON IMPROVING FEATURE SELECTION

Q: How can I download an article?

To download an article from SID, first log in to the site, search for the article title, and click on the 'Download Article' option.

Q: How can I download an ISI article?

To download an ISI article on SID, enter the keyword or article title in the search bar, view the relevant results, click on the desired article, and select the 'Download Article' option.

Q: How can I access the SID database?

To access the SID database, visit SID.ir, create an account, and log in to access scientific resources.

Q: Is downloading articles from SID free?

Some articles on SID are available for free, while others require payment. Details are specified on the article's page.

JALILI S.; BITARAFAN M.

Scientific Information Database (SID) - Trusted Source for Research and Academic Resources

Journal Paper

Paper Information

Journal: Year:2006 | Volume:40 | Issue:3 (97) Page(s): 313-328

Download Full-Text

مرکز اطلاعات علمی Scientific Information Database (SID) - Trusted Source for Research and Academic Resources

Persian Verion

View:

999

Download:

Cites:

Information Journal Paper

Title

INCREASING PERFORMANCE OF TEXT CLASSIFICATION BASED ON IMPROVING FEATURE SELECTION

Author(s)

JALILI S. | BITARAFAN M. | Issue Writer Certificate

Keywords

Not Registered.

Abstract

In text classification, words are usually used as features. Since text classification methods are faced to too many features, here we propose several methods to improve feature selection in text classification, and we show merits of the proposed methods in comparison to the relevant ones. First, we propose two methods that just concentrate on all positive and negative correlation factors between features and document categories. The first method uses all correlation factors with a positive effect while the second one uses positive factors with a positive effect and negative factors with a negative effect in feature selection criteria. Our experiments show that the positive factors are more effective than the negative factors in feature selection criteria. Next, we propose a hybrid method. First, it uses Relief-F filtering method to select a number of features spending a low computation cost. Then, the selected features are further reduced using an SFS or SBS wrapper method. The hybrid method has a better performance in comparison to filtering methods. Our experiment results show that applying this method to SVM Light classifier on Reuters-21578 corpora we can remove up to 94% of features while improving classification accuracy. Moreover, it is worth to mention that Relief-F by itself has shown a good result where it is the first time it is applied to text domain.

Cites

No record.

References

No record.

Cite

APA: Copy

JALILI, S., & BITARAFAN, M.. (2006). INCREASING PERFORMANCE OF TEXT CLASSIFICATION BASED ON IMPROVING FEATURE SELECTION. JOURNAL OF FACULTY OF ENGINEERING (UNIVERSITY OF TEHRAN), 40(3 (97)), 313-328. SID. https://sid.ir/paper/14467/en

Vancouver: Copy

JALILI S., BITARAFAN M.. INCREASING PERFORMANCE OF TEXT CLASSIFICATION BASED ON IMPROVING FEATURE SELECTION. JOURNAL OF FACULTY OF ENGINEERING (UNIVERSITY OF TEHRAN)[Internet]. 2006;40(3 (97)):313-328. Available from: https://sid.ir/paper/14467/en

IEEE: Copy

S. JALILI, and M. BITARAFAN, “INCREASING PERFORMANCE OF TEXT CLASSIFICATION BASED ON IMPROVING FEATURE SELECTION,” JOURNAL OF FACULTY OF ENGINEERING (UNIVERSITY OF TEHRAN), vol. 40, no. 3 (97), pp. 313–328, 2006, [Online]. Available: https://sid.ir/paper/14467/en

Related Journal Papers