To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Categorize the Opinions of Individuals using the Classifier Combination based on the Feature Selection |
|
Author |
Leila Delikhoon |
| Citation |
Vol. 26 No. 8 pp. 124-130 |
|
Abstract |
Today, data mining and information analysis attract attention of many researchers. Also , cultural, social, social, and commercial motivations increase the importance of the data mining. Since the user', behavior is very decisive, analysis the user', opinion provides very useful information to managers. In this study, a new method is introduced to categorize the views and opinions of the users by using the classifier combination and conduct voting . This method takes more time for learning, but increases the classification precision . Since, the number of features of the data set is very high, in order to compensate the speed drop, a method is selected for feature selection . This method once is performed for the trial data , and then classification is performed just by the extracted features that compensates the speed reduction of the classifier combination. To test the suggested method, three data sets of the users' view from the Amazon website called Book, DVD, and Kitchen, have been used. The results of the classification were evaluated by the total accuracy, precision , and recall criteria. These criteria show that, in addition to reduce the feature, algorithm of feature selection, increases the classification efficiency that it is due to remove the noise and inappropriate features for classification. Also , results show that classifier combination is more effective than the classification by one base algorithm. |
|
Keywords |
Data mining, classification, classifier combination, feature selection |
|
URL |
