Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255105
Hong Li
This paper, based on the author's new model of granular computing — the theory of granular set, offers firstly the definition of granule, and describes the form of four-tuple array of granule, then puts forward the concept of object granular set, description granular set, granular set, object granular system, description granular system, granular system, through upgrading the mapping from the Point Set to the Power Set and mapping from one-way to two-way. It further describes those concept s respectively, of which the description of granular set and granular system are in the form of five-tuple array, that is, (U, D, L, H, J), where U is the universe of the problem discussed, D describes all the elements in U, L and H are the operators of the opposite direction, and J restricts the L and H. The difference between granular set and granular system is that the operator in the granular set is from the Point Set to the Point Set, while the operator in the granular system is from the Power Set to the Power Set. Thus it expands and improves the granular set theory.
{"title":"Granule, granular set and granular system","authors":"Hong Li","doi":"10.1109/GRC.2009.5255105","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255105","url":null,"abstract":"This paper, based on the author's new model of granular computing — the theory of granular set, offers firstly the definition of granule, and describes the form of four-tuple array of granule, then puts forward the concept of object granular set, description granular set, granular set, object granular system, description granular system, granular system, through upgrading the mapping from the Point Set to the Power Set and mapping from one-way to two-way. It further describes those concept s respectively, of which the description of granular set and granular system are in the form of five-tuple array, that is, (U, D, L, H, J), where U is the universe of the problem discussed, D describes all the elements in U, L and H are the operators of the opposite direction, and J restricts the L and H. The difference between granular set and granular system is that the operator in the granular set is from the Point Set to the Point Set, while the operator in the granular system is from the Power Set to the Power Set. Thus it expands and improves the granular set theory.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127669564","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255081
X. Mu, Yaling Zhang
A improved gradient-based backpropagation training method is proposed for neural networks in this paper. Based on the Barzilai and Borwein steplength update and some technique of Resilient Propagation method, we adapt the new learning rate to improves the speed and the success rate. Experimental results show that the proposed method has considerably improved convergence speed, and for the chosen test problems, outperforms other well-known training methods.
{"title":"A modified gradient-based backpropagation training method for neural networks","authors":"X. Mu, Yaling Zhang","doi":"10.1109/GRC.2009.5255081","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255081","url":null,"abstract":"A improved gradient-based backpropagation training method is proposed for neural networks in this paper. Based on the Barzilai and Borwein steplength update and some technique of Resilient Propagation method, we adapt the new learning rate to improves the speed and the success rate. Experimental results show that the proposed method has considerably improved convergence speed, and for the chosen test problems, outperforms other well-known training methods.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129942993","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255062
Tetsuya Toyota, H. Nobuhara
In order to grasp the perspective of more than seven thousand laws in Japan, and to find the relationships between law and laws, a method of creating a hierarchical network of the laws by using the morphological analysis and granular computing, is proposed. The proposed method analyzes the hierarchical networks by using the index of the network science such as degree distribution. Furthermore, it visualizes the hierarchical structure in the setting of granular computing. By using JAVA-based language ‘Processing’, a network visualization system is implemented, and it is confirmed that users can easily analyze/understand the law network structure by the proposed system.
{"title":"Hierarchical structure analysis and visualization of Japanese law networks based on morphological analysis and granular computing","authors":"Tetsuya Toyota, H. Nobuhara","doi":"10.1109/GRC.2009.5255062","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255062","url":null,"abstract":"In order to grasp the perspective of more than seven thousand laws in Japan, and to find the relationships between law and laws, a method of creating a hierarchical network of the laws by using the morphological analysis and granular computing, is proposed. The proposed method analyzes the hierarchical networks by using the index of the network science such as degree distribution. Furthermore, it visualizes the hierarchical structure in the setting of granular computing. By using JAVA-based language ‘Processing’, a network visualization system is implemented, and it is confirmed that users can easily analyze/understand the law network structure by the proposed system.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129959273","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}
{"title":"Robust stability of Fuzzy Elman Neural Network","authors":"L. F. Araghi, H. Shah-Hosseini","doi":"10.1109/GRC.2009.5255173","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255173","url":null,"abstract":"This paper proposed three methods for existence of a common quadratic Lyapunov function for Robust Stability Analysis of Fuzzy Elman Neural Network.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129881031","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255169
S. Bai, Sixue Bai
The DNA sequence data is one of the basic and important data among biological data. The DNA sequence pattern mining has got wide attention and rapid development. Traditional algorithms for the sequential pattern mining may generate lots of redundant patterns when dealing with the DNA sequence. The Maximal Frequent Pattern is preferable to express the function and structure of the DNA sequence. Base on the characteristics of the DNA sequence, the author develops the Joined Maximal Pattern Segments algorithm—JMPS, for the maximal frequent patterns mining of the DNA sequence. First, the maximal frequent pattern segments base on adjacent generated. Then, longer Maximal Frequent Pattern can be obtained by combining the above segments, at the same time deleting the Non-maximal patterns. The algorithm can deal with the DNA sequence data efficiently.
{"title":"The Maximal Frequent Pattern mining of DNA sequence","authors":"S. Bai, Sixue Bai","doi":"10.1109/GRC.2009.5255169","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255169","url":null,"abstract":"The DNA sequence data is one of the basic and important data among biological data. The DNA sequence pattern mining has got wide attention and rapid development. Traditional algorithms for the sequential pattern mining may generate lots of redundant patterns when dealing with the DNA sequence. The Maximal Frequent Pattern is preferable to express the function and structure of the DNA sequence. Base on the characteristics of the DNA sequence, the author develops the Joined Maximal Pattern Segments algorithm—JMPS, for the maximal frequent patterns mining of the DNA sequence. First, the maximal frequent pattern segments base on adjacent generated. Then, longer Maximal Frequent Pattern can be obtained by combining the above segments, at the same time deleting the Non-maximal patterns. The algorithm can deal with the DNA sequence data efficiently.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117070855","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255138
Liang Gao, Shaoyue Yu, Yu-Pan Luo, L. Shang
Outlier detection is an important data mining task, LOF(local outlier factor) was proposed to indicate the degree of outlier-ness, which is practical for finding local outliers. However, it is difficult to decide the neighborhood size. In this paper a multi-granularity local outlier detection(MLOD) method is proposed to organize the outlierness under multi-granularity. It finds local outliers in varying neighborhood granularity. This method applies approximation as well as grid-based partition to reduce time complexity. The theoretical results show that the time cost is linear to the size of data sets. Furthermore, the provided output and analysis can also assist users to choose the appropriate parameters. The performance of the algorithm is presented by experimenting on three generated data sets.
{"title":"MLOD: Multi-granularity local outlier detection","authors":"Liang Gao, Shaoyue Yu, Yu-Pan Luo, L. Shang","doi":"10.1109/GRC.2009.5255138","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255138","url":null,"abstract":"Outlier detection is an important data mining task, LOF(local outlier factor) was proposed to indicate the degree of outlier-ness, which is practical for finding local outliers. However, it is difficult to decide the neighborhood size. In this paper a multi-granularity local outlier detection(MLOD) method is proposed to organize the outlierness under multi-granularity. It finds local outliers in varying neighborhood granularity. This method applies approximation as well as grid-based partition to reduce time complexity. The theoretical results show that the time cost is linear to the size of data sets. Furthermore, the provided output and analysis can also assist users to choose the appropriate parameters. The performance of the algorithm is presented by experimenting on three generated data sets.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122670697","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255153
Tzung-Nan Chuang, Chia-Tzu Lin, J. Kung, Ming-Da Lin
Nowadays, liner shipping has become a constant operation model for shipping companies, and scheduling is an important issue for operation. It is well-known that a nice plan for route of container ships will bring long-term profit to companies. In the earlier works, the market demand is assumed to be crisp. However, the market demand could be uncertain in real world. Fuzzy sets theory is frequently used to deal with the uncertainty problem. On the other hand, genetic algorithm owns powerful multi-objective searching capability and it can extensively find optimal solutions through continuous copy, crossover, and mutation. Due to these advantages, in this paper, a fuzzy genetic algorithm for liner shipping planning is proposed. This algorithm not only takes market demand, shipping and berthing time of container ships into account simultaneously but also is capable of finding the most suitable route of container ships.
{"title":"Fuzzy genetic algorithm for the route of container ships","authors":"Tzung-Nan Chuang, Chia-Tzu Lin, J. Kung, Ming-Da Lin","doi":"10.1109/GRC.2009.5255153","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255153","url":null,"abstract":"Nowadays, liner shipping has become a constant operation model for shipping companies, and scheduling is an important issue for operation. It is well-known that a nice plan for route of container ships will bring long-term profit to companies. In the earlier works, the market demand is assumed to be crisp. However, the market demand could be uncertain in real world. Fuzzy sets theory is frequently used to deal with the uncertainty problem. On the other hand, genetic algorithm owns powerful multi-objective searching capability and it can extensively find optimal solutions through continuous copy, crossover, and mutation. Due to these advantages, in this paper, a fuzzy genetic algorithm for liner shipping planning is proposed. This algorithm not only takes market demand, shipping and berthing time of container ships into account simultaneously but also is capable of finding the most suitable route of container ships.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"489 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123027640","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255139
Xuewen Chen
While much of molecular biology research has led to a wealth of knowledge about individual cellular components and their functions, it has become increasingly clear that most cellular functions are carried out by complex networks of interconnected components, and that the characterization of isolated cellular components is not sufficient to understand the cell's complexity. In recent years, the development of high-throughput technologies has provided the scientific community with exciting new opportunities for systematically studying biological networks on a whole-genome scale. One of the great challenges currently confronting scientists in systems biology research is how to computationally model and elucidate the function and the mechanisms of the complex biological networks from these high-throughput biological data sets. In this talk, I will discuss some machine learning methods recently developed in my group for uncovering genes involved in the same pathways and for predicting protein-protein interactions and protein functions.
{"title":"Computational models in systems biology","authors":"Xuewen Chen","doi":"10.1109/GRC.2009.5255139","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255139","url":null,"abstract":"While much of molecular biology research has led to a wealth of knowledge about individual cellular components and their functions, it has become increasingly clear that most cellular functions are carried out by complex networks of interconnected components, and that the characterization of isolated cellular components is not sufficient to understand the cell's complexity. In recent years, the development of high-throughput technologies has provided the scientific community with exciting new opportunities for systematically studying biological networks on a whole-genome scale. One of the great challenges currently confronting scientists in systems biology research is how to computationally model and elucidate the function and the mechanisms of the complex biological networks from these high-throughput biological data sets. In this talk, I will discuss some machine learning methods recently developed in my group for uncovering genes involved in the same pathways and for predicting protein-protein interactions and protein functions.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130800881","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255021
Feng Zhan, Keming Xie, Jingge Zhao, Gang Xie
The hierarchical fault diagnosis based on granular matrix and Signed Directed Graph (SDG) is presented in the paper. Granular Computing (GrC) theory can be introduced into SDG-based fault diagnosis to optimize the decision table. The rules of fault diagnosis are reasoned out through searching the associated path of the SDG model. The redundant nodes of the failure diagnosis rules are reduced by the attribute reduction algorithm based on granular matrix, which can simplify the solution of failure diagnosis, avoid the setting of the redundant sensor, and decrease the complexity of collocating sensor network. Compared with the traditional failure diagnosis based on SDG, the designed scheme and an experimental example of a hot nitric acid cooling failure diagnosis system show that the hierarchical fault diagnosis based on granular matrix and SDG in the paper is not only feasibly and effectively, but also valuable in practice.
{"title":"Fault diagnosis based on granular matrix-SDG and its application","authors":"Feng Zhan, Keming Xie, Jingge Zhao, Gang Xie","doi":"10.1109/GRC.2009.5255021","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255021","url":null,"abstract":"The hierarchical fault diagnosis based on granular matrix and Signed Directed Graph (SDG) is presented in the paper. Granular Computing (GrC) theory can be introduced into SDG-based fault diagnosis to optimize the decision table. The rules of fault diagnosis are reasoned out through searching the associated path of the SDG model. The redundant nodes of the failure diagnosis rules are reduced by the attribute reduction algorithm based on granular matrix, which can simplify the solution of failure diagnosis, avoid the setting of the redundant sensor, and decrease the complexity of collocating sensor network. Compared with the traditional failure diagnosis based on SDG, the designed scheme and an experimental example of a hot nitric acid cooling failure diagnosis system show that the hierarchical fault diagnosis based on granular matrix and SDG in the paper is not only feasibly and effectively, but also valuable in practice.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126641539","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}
Pub Date : 2009-09-22DOI: 10.1109/GRC.2009.5255017
Weiwei Zhang, Wei Chen, Zhenyuan Wang
The Choquet integral has been applied in data mining, such as nonlinear multiregressions and nonlinear classifications. Adopting signed efficiency measures in the Choquet integral makes the models more powerful. Another idea for generalizing the above-mensioned models is to use a linear core in the Choquet integral. This has been successfully used in nonlinear mulregression. However, there is a uniqueness problem for presenting the Choquet integral in classification models such that it is difficult to explain the exact contribution rate from each individual attributes, as well as their combinations, towards the target. In this work, an additional restriction on the parameters is given to guarantee the uniqueness of the expression.
{"title":"On the uniqueness of the expression for the Choquet integral with linear core in classification","authors":"Weiwei Zhang, Wei Chen, Zhenyuan Wang","doi":"10.1109/GRC.2009.5255017","DOIUrl":"https://doi.org/10.1109/GRC.2009.5255017","url":null,"abstract":"The Choquet integral has been applied in data mining, such as nonlinear multiregressions and nonlinear classifications. Adopting signed efficiency measures in the Choquet integral makes the models more powerful. Another idea for generalizing the above-mensioned models is to use a linear core in the Choquet integral. This has been successfully used in nonlinear mulregression. However, there is a uniqueness problem for presenting the Choquet integral in classification models such that it is difficult to explain the exact contribution rate from each individual attributes, as well as their combinations, towards the target. In this work, an additional restriction on the parameters is given to guarantee the uniqueness of the expression.","PeriodicalId":388774,"journal":{"name":"2009 IEEE International Conference on Granular Computing","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116984770","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：