Automatic text classification has gained huge popularity with the advancement of information technology. Bayesian method has been found highly appropriate for text classification but it suffers from a number of problems. When there is large number of categories, lack of uniformity in training data becomes a big problem. Some nodes may get less training documents, while other may get a very large number. Therefore, some nodes are biased over others. Besides, presence of noise data or outliers also creates problems. Moreover, when documents are very small, just like a line item describing a product, the problem becomes more difficult. In this paper we describe a method that combines naive Bayesian text classification technique and neural networks to handle these problems. We start with a naive Bayesian classifier, which has the linear separating surfaces. We modify the separating surfaces using neural network to find better separating surfaces and hence better classification accuracy over validation data.
{"title":"Knowledge Based Neural Network for Text Classification","authors":"R. D. Goyal","doi":"10.1109/GrC.2007.108","DOIUrl":"https://doi.org/10.1109/GrC.2007.108","url":null,"abstract":"Automatic text classification has gained huge popularity with the advancement of information technology. Bayesian method has been found highly appropriate for text classification but it suffers from a number of problems. When there is large number of categories, lack of uniformity in training data becomes a big problem. Some nodes may get less training documents, while other may get a very large number. Therefore, some nodes are biased over others. Besides, presence of noise data or outliers also creates problems. Moreover, when documents are very small, just like a line item describing a product, the problem becomes more difficult. In this paper we describe a method that combines naive Bayesian text classification technique and neural networks to handle these problems. We start with a naive Bayesian classifier, which has the linear separating surfaces. We modify the separating surfaces using neural network to find better separating surfaces and hence better classification accuracy over validation data.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132023799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ontology evaluation is one of the most important phases of ontology engineering. Researchers have identified different types of errors that should be catered in ontology evaluation process for fulfillment of the semantic Web vision and classified them in error's taxonomy. We have found that some important errors are missing in the error's taxonomy. We have identified and defined two new incompleteness errors i.e. functional property omission (FPO) for single valued property and inverse-functional property omission (IFPO) for a unique valued property. We have demonstrated the importance of such errors by giving different scenarios where appropriate. We have evaluated different ontologies and presented empirical results.
{"title":"Incompleteness Errors in Ontology","authors":"M. Qadir, Muhammad Fahad, Syed Adnan Hussain Shah","doi":"10.1109/GrC.2007.152","DOIUrl":"https://doi.org/10.1109/GrC.2007.152","url":null,"abstract":"Ontology evaluation is one of the most important phases of ontology engineering. Researchers have identified different types of errors that should be catered in ontology evaluation process for fulfillment of the semantic Web vision and classified them in error's taxonomy. We have found that some important errors are missing in the error's taxonomy. We have identified and defined two new incompleteness errors i.e. functional property omission (FPO) for single valued property and inverse-functional property omission (IFPO) for a unique valued property. We have demonstrated the importance of such errors by giving different scenarios where appropriate. We have evaluated different ontologies and presented empirical results.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132981224","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Detecting unaffected races is important for debugging MPI parallel programs, because unaffected races can cause the occurrence of affected races which do not need to be debugged. However, the previous techniques can not discern unaffected races from affected races so that programmers will be easily overwhelmed by the vast information of race detection. In this paper, we present a new visualization which lets programmers know which race is affected or not. For this, our technique checks whether any message racing toward a race is affected or not based on happen- before relation, and also checks which process influences a race during an execution. After the execution, it visualizes the affect-relations of the detected races. Therefore, our visualization helps for programmers to effectively distinguish unaffected races from affected races, and to debug MPI parallel programs.
{"title":"Visualization of Affect-Relations of Message Races for Debugging MPI Programs","authors":"Mi-Young Park, S. Kim, Hyuk-Ro Park","doi":"10.1109/GrC.2007.120","DOIUrl":"https://doi.org/10.1109/GrC.2007.120","url":null,"abstract":"Detecting unaffected races is important for debugging MPI parallel programs, because unaffected races can cause the occurrence of affected races which do not need to be debugged. However, the previous techniques can not discern unaffected races from affected races so that programmers will be easily overwhelmed by the vast information of race detection. In this paper, we present a new visualization which lets programmers know which race is affected or not. For this, our technique checks whether any message racing toward a race is affected or not based on happen- before relation, and also checks which process influences a race during an execution. After the execution, it visualizes the affect-relations of the detected races. Therefore, our visualization helps for programmers to effectively distinguish unaffected races from affected races, and to debug MPI parallel programs.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133189494","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper analyzes the global asymptotic stability of delayed Hopfield neural networks by utilizing Lyapunov functional method and a generalized inequality technique. A new sufficient condition ensuring global asymptotic stability of the unique equilibrium point of delayed Hopfield neural networks is obtained. The result is related to the size of delays. The obtained conditions show to be less conservative and restrictive than that reported in the literature. A numerical simulation is given to illustrate the efficiency of our result.
{"title":"A Novel Delay-Dependent Global Stability Criterion of Delayed Hopfield Neural Networks","authors":"Degang Yang, Qun Liu, Yong Wang","doi":"10.1109/GrC.2007.16","DOIUrl":"https://doi.org/10.1109/GrC.2007.16","url":null,"abstract":"This paper analyzes the global asymptotic stability of delayed Hopfield neural networks by utilizing Lyapunov functional method and a generalized inequality technique. A new sufficient condition ensuring global asymptotic stability of the unique equilibrium point of delayed Hopfield neural networks is obtained. The result is related to the size of delays. The obtained conditions show to be less conservative and restrictive than that reported in the literature. A numerical simulation is given to illustrate the efficiency of our result.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130059602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Text classification algorithms, such SVM, and Naive Bayes, have been developed to build up search engines and construct spam email filters. As a simple yet powerful sample of Bayesian theorem, naive Bayes shows advantages in text classification yielding satisfactory results. In this paper, a spam email detector is developed using naive Bayes algorithm. We use pre-classified emails (priory knowledge) to train the spam email detector. With the model generated from the training step, the detector is able to decide whether an email is a spam email or an ordinary email.
{"title":"Naïve Bayes Text Classifier","authors":"Haiyi Zhang, Di Li","doi":"10.1109/GrC.2007.40","DOIUrl":"https://doi.org/10.1109/GrC.2007.40","url":null,"abstract":"Text classification algorithms, such SVM, and Naive Bayes, have been developed to build up search engines and construct spam email filters. As a simple yet powerful sample of Bayesian theorem, naive Bayes shows advantages in text classification yielding satisfactory results. In this paper, a spam email detector is developed using naive Bayes algorithm. We use pre-classified emails (priory knowledge) to train the spam email detector. With the model generated from the training step, the detector is able to decide whether an email is a spam email or an ordinary email.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121371339","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Social network analysis is a methodology used extensively in social and behavioral sciences, as well as in political science, economics, organization theory, and industrial engineering. Positional analysis of a social network aims to find similarities between actors in the network. One of the the most studied notions in the positional analysis of social networks is regular equivalence. According to Borgatti and Everett, two actors are regularly equivalent if they are equally related to equivalent others. In recent years, fuzzy social networks have also received considerable attention because they can represent both the qualitative relationship and the degrees of interaction between actors. In this paper, we generalize the notion of regular equivalence to fuzzy social networks based on two alternative definitions of regular equivalence. While these two definitions are equivalent for social networks, they induce different generalizations for fuzzy social networks. The first generalization, called regular similarity, is based on the characterization of regular equivalence as an equivalence relation that commutes with the underlying social relations. The regular similarity is then a fuzzy binary relation that specifies the degree of similarity between actors in the social network. The second generalization, called generalized regular equivalence, is based on the definition of role assignment or coloring. A role assignment (resp. coloring) is a mapping from the set of actors to a set of roles (resp. colors). The mapping is regular if actors assigned to the same role have the same roles in their neighborhoods. Consequently, generalized regular equivalence is an equivalence relation that can determine the role partition of the actors in a fuzzy social network.
{"title":"Positional Analysis in Fuzzy Social Networks","authors":"T. Fan, C. Liau, T. Lin","doi":"10.1109/GrC.2007.9","DOIUrl":"https://doi.org/10.1109/GrC.2007.9","url":null,"abstract":"Social network analysis is a methodology used extensively in social and behavioral sciences, as well as in political science, economics, organization theory, and industrial engineering. Positional analysis of a social network aims to find similarities between actors in the network. One of the the most studied notions in the positional analysis of social networks is regular equivalence. According to Borgatti and Everett, two actors are regularly equivalent if they are equally related to equivalent others. In recent years, fuzzy social networks have also received considerable attention because they can represent both the qualitative relationship and the degrees of interaction between actors. In this paper, we generalize the notion of regular equivalence to fuzzy social networks based on two alternative definitions of regular equivalence. While these two definitions are equivalent for social networks, they induce different generalizations for fuzzy social networks. The first generalization, called regular similarity, is based on the characterization of regular equivalence as an equivalence relation that commutes with the underlying social relations. The regular similarity is then a fuzzy binary relation that specifies the degree of similarity between actors in the social network. The second generalization, called generalized regular equivalence, is based on the definition of role assignment or coloring. A role assignment (resp. coloring) is a mapping from the set of actors to a set of roles (resp. colors). The mapping is regular if actors assigned to the same role have the same roles in their neighborhoods. Consequently, generalized regular equivalence is an equivalence relation that can determine the role partition of the actors in a fuzzy social network.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124578868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We propose an anaphor resolution based opinion holder identification method exploiting lexical and syntactic information. We tested our approach on online news documents and obtained 72.22% and 69.89% in accuracy for the task of non-anaphoric opinion holder resolution and the task of anaphoric opinion holder identification, respectively.
{"title":"Identifying Opinion Holders in Opinion Text from Online Newspapers","authors":"Youngho Kim, Yuchul Jung, Sung-Hyon Myaeng","doi":"10.1109/GRC.2007.82","DOIUrl":"https://doi.org/10.1109/GRC.2007.82","url":null,"abstract":"We propose an anaphor resolution based opinion holder identification method exploiting lexical and syntactic information. We tested our approach on online news documents and obtained 72.22% and 69.89% in accuracy for the task of non-anaphoric opinion holder resolution and the task of anaphoric opinion holder identification, respectively.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117184371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Syoji Kobashi, Y. Fujimoto, M. Ogawa, K. Ando, R. Ishikura, K. Kondo, S. Hirota, Y. Hata
Automated stripping of skulls from infantile brain MR images is the fundamental work to visualize cerebral surface and to measure cerebral volumes. They are important to evaluate cerebral diseases because most cerebral diseases cause morphometric changes in cerebrum. This study proposes a novel image segmentation method based on fuzzy rule-based active surface model. The proposed method was validated by applying it to two neonatal (3W and 4W) and six infantile (5W to 4Y2M) subjects. The mean sensitivity was 98.84 %, and false-positive rate was 1.21 %, and the cerebral surface was visualized well.
{"title":"Fuzzy-ASM Based Automated Skull Stripping Method from Infantile Brain MR Images","authors":"Syoji Kobashi, Y. Fujimoto, M. Ogawa, K. Ando, R. Ishikura, K. Kondo, S. Hirota, Y. Hata","doi":"10.1109/GrC.2007.63","DOIUrl":"https://doi.org/10.1109/GrC.2007.63","url":null,"abstract":"Automated stripping of skulls from infantile brain MR images is the fundamental work to visualize cerebral surface and to measure cerebral volumes. They are important to evaluate cerebral diseases because most cerebral diseases cause morphometric changes in cerebrum. This study proposes a novel image segmentation method based on fuzzy rule-based active surface model. The proposed method was validated by applying it to two neonatal (3W and 4W) and six infantile (5W to 4Y2M) subjects. The mean sensitivity was 98.84 %, and false-positive rate was 1.21 %, and the cerebral surface was visualized well.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129546891","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The literature exploits an intelligent method to spread rough sets to normed linear space where there is a basis, establishes rough sets in normed linear space, and puts forward an upper and lower approximation calculation formula. This paper further researched the problems of rough sets in normed space and obtained some useful results.
{"title":"Properties of Rough Sets in Normed Linear Space and its Proof","authors":"Hui Sun, Y. Wang, L. He, Qing Liu","doi":"10.1109/GrC.2007.111","DOIUrl":"https://doi.org/10.1109/GrC.2007.111","url":null,"abstract":"The literature exploits an intelligent method to spread rough sets to normed linear space where there is a basis, establishes rough sets in normed linear space, and puts forward an upper and lower approximation calculation formula. This paper further researched the problems of rough sets in normed space and obtained some useful results.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129697331","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we propose two techniques for face recognition, namely, view-based and biometric-based face recognition. Both use general backpropagation neural networks (GBPN's). In the view-based method, we extract sub-images of the eyes, the nose, and the mouth and feed them into a GBPN. In the biometric-based method, seven measurements of the face will be fed into another GBPN. We illustrate the results of the proposed algorithms by applying them on the Cambridge ORL face database, which contains quite a high degree of variability in expression, pose, and facial details. We have found that the view-based method outperforms the biometric-based method. Thus, we have selected the view-based method to function as the main neural network whereas the biometric-based method will function as a supportive neural network.
{"title":"Face Recognition Using a Hybrid General Backpropagation Neural Network","authors":"M. S. Charifa, A. Suliman, M. Bikdash","doi":"10.1109/GrC.2007.136","DOIUrl":"https://doi.org/10.1109/GrC.2007.136","url":null,"abstract":"In this paper, we propose two techniques for face recognition, namely, view-based and biometric-based face recognition. Both use general backpropagation neural networks (GBPN's). In the view-based method, we extract sub-images of the eyes, the nose, and the mouth and feed them into a GBPN. In the biometric-based method, seven measurements of the face will be fed into another GBPN. We illustrate the results of the proposed algorithms by applying them on the Cambridge ORL face database, which contains quite a high degree of variability in expression, pose, and facial details. We have found that the view-based method outperforms the biometric-based method. Thus, we have selected the view-based method to function as the main neural network whereas the biometric-based method will function as a supportive neural network.","PeriodicalId":259430,"journal":{"name":"2007 IEEE International Conference on Granular Computing (GRC 2007)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121483811","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：