Support vector machines (SVM) and relevance vector machines (RVM) constitute two state-of-the-art learning machines that are currently focus of cutting-edge research. SVM present accuracy and complexity preponderance, but are surpassed by RVM when probabilistic outputs or kernel selection come to discussion. We propose a two-level hierarchical hybrid SVM-RVM model to combine the best of both learning machines. The proposed model first level uses an RVM to determine the less confident classified examples and the second level then makes use of an SVM to learn and classify the tougher examples. We show the benefits of the hierarchical approach on a text classification task, where the two-levels outperform both learning machines
{"title":"Two-Level Hierarchical Hybrid SVM-RVM Classification Model","authors":"Catarina Silva, B. Ribeiro","doi":"10.1109/ICMLA.2006.52","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.52","url":null,"abstract":"Support vector machines (SVM) and relevance vector machines (RVM) constitute two state-of-the-art learning machines that are currently focus of cutting-edge research. SVM present accuracy and complexity preponderance, but are surpassed by RVM when probabilistic outputs or kernel selection come to discussion. We propose a two-level hierarchical hybrid SVM-RVM model to combine the best of both learning machines. The proposed model first level uses an RVM to determine the less confident classified examples and the second level then makes use of an SVM to learn and classify the tougher examples. We show the benefits of the hierarchical approach on a text classification task, where the two-levels outperform both learning machines","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"101 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115766846","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 present an automatic spatio-temporal mining system of rolling and adherent leukocytes for intravital videos. The magnitude of leukocyte adhesion and the decrease in rolling velocity are common interests for inflammation response studies. Currently, there is no existing system which is perfect for such purposes. Our approach starts with locating moving leukocytes by probabilistic learning of temporal features. It then removes noises through median and location-based filtering, and finally performs motion correspondence through centroid trackers. By extracting the information about moving leukocytes first, we are able to extract adherent leukocytes in a more robust way with an adaptive threshold method. The effectiveness and the efficiency of the proposed method are demonstrated by the experimental results
{"title":"Automatic Intravital Video Mining of Rolling and Adhering Leukocytes","authors":"Xin C. Anders, Chengcui Zhang, Hong Yuan","doi":"10.1109/ICMLA.2006.18","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.18","url":null,"abstract":"In this paper, we present an automatic spatio-temporal mining system of rolling and adherent leukocytes for intravital videos. The magnitude of leukocyte adhesion and the decrease in rolling velocity are common interests for inflammation response studies. Currently, there is no existing system which is perfect for such purposes. Our approach starts with locating moving leukocytes by probabilistic learning of temporal features. It then removes noises through median and location-based filtering, and finally performs motion correspondence through centroid trackers. By extracting the information about moving leukocytes first, we are able to extract adherent leukocytes in a more robust way with an adaptive threshold method. The effectiveness and the efficiency of the proposed method are demonstrated by the experimental results","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115772552","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}
Modularity and building blocks have drawn attention from the genetic programming (GP) community for a long time. The results are usually twofold: a hierarchical evolution with adequate building block reuse can accelerate the learning process, but rigidly defined and excessively employed modules may also counteract the expected advantages by confining the reachable search space. In this work, we introduce the concept of emergent loose modules based on a new linear GP system, prefix gene expression programming (P-GEP), in an attempt to balance between the stochastic exploration and the hierarchical construction for the optimal solutions. Emergent loose modules are dynamically produced by the evolution, and are reusable as sub-functions in later generations. The proposed technique is fully illustrated with a simple symbolic regression problem. The initial experimental results suggest it is a flexible approach in identifying the evolved regularity and the emergent loose modules are critical in composing the best solutions
{"title":"Introducing Emergent Loose Modules into the Learning Process of a Linear Genetic Programming System","authors":"Xin Li, Chi Zhou, Weimin Xiao, P. Nelson","doi":"10.1109/ICMLA.2006.31","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.31","url":null,"abstract":"Modularity and building blocks have drawn attention from the genetic programming (GP) community for a long time. The results are usually twofold: a hierarchical evolution with adequate building block reuse can accelerate the learning process, but rigidly defined and excessively employed modules may also counteract the expected advantages by confining the reachable search space. In this work, we introduce the concept of emergent loose modules based on a new linear GP system, prefix gene expression programming (P-GEP), in an attempt to balance between the stochastic exploration and the hierarchical construction for the optimal solutions. Emergent loose modules are dynamically produced by the evolution, and are reusable as sub-functions in later generations. The proposed technique is fully illustrated with a simple symbolic regression problem. The initial experimental results suggest it is a flexible approach in identifying the evolved regularity and the emergent loose modules are critical in composing the best solutions","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130487156","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}
Tree induction is one of the most effective and widely used models in classification. Unfortunately, decision trees such as C4.5 have been found to provide poor probability estimates. By the empirical studies, Provost and Domingos found that probability estimation trees (PETs) give a fairly good probability estimation. However, different from normal decision trees, pruning reduces the performances of PETs. In order to get a good probability estimation, we usually need large trees which are not good in terms of the model transparency. In this paper, two hybrid models by combining the naive Bayes classifier and PETs are proposed in order to build a model with good performance without losing too much transparency. The first model use naive Bayes estimation given a PET and the second model use a group of small-sized PETs as naive Bayes estimators. Empirical studies show that the first model outperforms the PET model at shallow depth and the second model is equivalent to naive Bayes and PET
{"title":"Naive Bayes Classification Given Probability Estimation Trees","authors":"Zengchang Qin","doi":"10.1109/ICMLA.2006.36","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.36","url":null,"abstract":"Tree induction is one of the most effective and widely used models in classification. Unfortunately, decision trees such as C4.5 have been found to provide poor probability estimates. By the empirical studies, Provost and Domingos found that probability estimation trees (PETs) give a fairly good probability estimation. However, different from normal decision trees, pruning reduces the performances of PETs. In order to get a good probability estimation, we usually need large trees which are not good in terms of the model transparency. In this paper, two hybrid models by combining the naive Bayes classifier and PETs are proposed in order to build a model with good performance without losing too much transparency. The first model use naive Bayes estimation given a PET and the second model use a group of small-sized PETs as naive Bayes estimators. Empirical studies show that the first model outperforms the PET model at shallow depth and the second model is equivalent to naive Bayes and PET","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115110964","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}
Binary support vector machines (SVM) have proven effective in classification. However, problems remain with respect to feature selection in multi-class classification. This article proposes a novel multi-class SVM, which performs classification and feature selection simultaneously via L1-norm penalized sparse representations. The proposed methodology, together with our developed regularization solution path, permits feature selection within the framework of classification. The operational characteristics of the proposed methodology is examined via both simulated and benchmark examples, and is compared to some competitors in terms of the accuracy of prediction and feature selection. The numerical results suggest that the proposed methodology is highly competitive
{"title":"On L_1-Norm Multi-class Support Vector Machines","authors":"Lifeng Wang, Xiaotong Shen, Yuan F. Zheng","doi":"10.1109/ICMLA.2006.38","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.38","url":null,"abstract":"Binary support vector machines (SVM) have proven effective in classification. However, problems remain with respect to feature selection in multi-class classification. This article proposes a novel multi-class SVM, which performs classification and feature selection simultaneously via L1-norm penalized sparse representations. The proposed methodology, together with our developed regularization solution path, permits feature selection within the framework of classification. The operational characteristics of the proposed methodology is examined via both simulated and benchmark examples, and is compared to some competitors in terms of the accuracy of prediction and feature selection. The numerical results suggest that the proposed methodology is highly competitive","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116252588","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 consider the problem of interactive iterative analysis of datasets that consist of a large number of records represented as feature vectors. The record set is known to contain a number of anomalous records that the analyst desires to locate and describe in a short and comprehensive manner The nature of the anomaly is not known in advance (in particular, it is not known, which features or feature values identify the anomalous records, and which are irrelevant to the search), and becomes clear only in the process of analysis, as the description of the target subset is gradually refined. This situation is common in computer intrusion analysis, when a forensic analyst browses the logs to locate traces of an intrusion of unknown nature and origin, and extends to other tasks and data sets. To facilitate such "browsing for anomalies", we propose an unsupervised data organization technique for initial summarization and representation of data sets, and a semi-supervised learning technique for iterative modifications of the latter representation. Our approach is based on information content and Jensen-Shannon divergence and is related to information bottleneck methods. We have implemented it as apart of the Kerf log analysis toolkit
{"title":"Semi-supervised Data Organization for Interactive Anomaly Analysis.","authors":"J. Aslam, S. Bratus, Virgil Pavlu","doi":"10.1109/ICMLA.2006.47","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.47","url":null,"abstract":"We consider the problem of interactive iterative analysis of datasets that consist of a large number of records represented as feature vectors. The record set is known to contain a number of anomalous records that the analyst desires to locate and describe in a short and comprehensive manner The nature of the anomaly is not known in advance (in particular, it is not known, which features or feature values identify the anomalous records, and which are irrelevant to the search), and becomes clear only in the process of analysis, as the description of the target subset is gradually refined. This situation is common in computer intrusion analysis, when a forensic analyst browses the logs to locate traces of an intrusion of unknown nature and origin, and extends to other tasks and data sets. To facilitate such \"browsing for anomalies\", we propose an unsupervised data organization technique for initial summarization and representation of data sets, and a semi-supervised learning technique for iterative modifications of the latter representation. Our approach is based on information content and Jensen-Shannon divergence and is related to information bottleneck methods. We have implemented it as apart of the Kerf log analysis toolkit","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122069405","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}
Support vector machine (SVM) is a powerful tool for binary classification. Numerous methods are known to fuse several binary SVMs into multi-class (MC) classifiers. These methods are efficient, but an accurate study of the misclassified items leads to notice two sources of mistakes: (1) the response of each classifier does not use the entire information from the SVM, and (2) the decision method does not use the entire information from the classifier responses. In this paper, we present a method which partially prevents these two losses of information by applying belief theories (BTs) to SVM fusion, while keeping the efficient aspect of the classical methods
{"title":"Modeling Hesitation and Conflict: A Belief-Based Approach for Multi-class Problems","authors":"Thomas Burger, O. Aran, A. Caplier","doi":"10.1109/ICMLA.2006.35","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.35","url":null,"abstract":"Support vector machine (SVM) is a powerful tool for binary classification. Numerous methods are known to fuse several binary SVMs into multi-class (MC) classifiers. These methods are efficient, but an accurate study of the misclassified items leads to notice two sources of mistakes: (1) the response of each classifier does not use the entire information from the SVM, and (2) the decision method does not use the entire information from the classifier responses. In this paper, we present a method which partially prevents these two losses of information by applying belief theories (BTs) to SVM fusion, while keeping the efficient aspect of the classical methods","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127360538","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}
Adaptivity in neural networks aims at equipping learning algorithms with the ability to self-update as new training data becomes available. In many application, data arrives over long periods of time, hence the traditional one-shot training phase cannot be applied. The most appropriate training methodology in such circumstances is incremental learning (IL). The present paper introduces a new IL algorithm dedicated to classification problems. The basic idea is to incrementally generate prototyped categories which are then linked to their corresponding classes. Numerical simulations show the performance of the proposed algorithm
{"title":"Incremental Learning By Decomposition","authors":"A. Bouchachia","doi":"10.1109/ICMLA.2006.28","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.28","url":null,"abstract":"Adaptivity in neural networks aims at equipping learning algorithms with the ability to self-update as new training data becomes available. In many application, data arrives over long periods of time, hence the traditional one-shot training phase cannot be applied. The most appropriate training methodology in such circumstances is incremental learning (IL). The present paper introduces a new IL algorithm dedicated to classification problems. The basic idea is to incrementally generate prototyped categories which are then linked to their corresponding classes. Numerical simulations show the performance of the proposed algorithm","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131337292","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}
Formal concept analysis (FCA) is an effective tool for data analysis and knowledge discovery. Concept lattice, which is derived from mathematical order theory and lattice theory, is the core of FCA. Many research works of various areas show that concept lattices structures is an effective platform for data mining, machine learning, information retrieval, software engineer, etc. This paper offers a brief overview of FCA and proposes to apply FCA as a tool for analysis and visualization of data in digital ecosystem, and also discusses the applications of data mining for digital ecosystem
{"title":"Formal Concept Analysis for Digital Ecosystem","authors":"Huaiguo Fu","doi":"10.1109/ICMLA.2006.24","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.24","url":null,"abstract":"Formal concept analysis (FCA) is an effective tool for data analysis and knowledge discovery. Concept lattice, which is derived from mathematical order theory and lattice theory, is the core of FCA. Many research works of various areas show that concept lattices structures is an effective platform for data mining, machine learning, information retrieval, software engineer, etc. This paper offers a brief overview of FCA and proposes to apply FCA as a tool for analysis and visualization of data in digital ecosystem, and also discusses the applications of data mining for digital ecosystem","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123345603","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}
Many feature selection models have been proposed for online handwriting recognition. However, most of them require expensive computational overhead, or inaccurately find an improper feature set which leads to unacceptable recognition rates. This paper presents a new efficient feature selection model for handwriting symbol recognition by using an improved sequential floating search method coupled with a hybrid classifier, which is obtained by combining hidden Markov models with multilayer forward network. The effectiveness of proposed method is verified by comprehensive experiments based on UNIPEN database
{"title":"A Fast Feature Selection Model for Online Handwriting Symbol Recognition","authors":"B. Huang, Mohand Tahar Kechadi","doi":"10.1109/ICMLA.2006.6","DOIUrl":"https://doi.org/10.1109/ICMLA.2006.6","url":null,"abstract":"Many feature selection models have been proposed for online handwriting recognition. However, most of them require expensive computational overhead, or inaccurately find an improper feature set which leads to unacceptable recognition rates. This paper presents a new efficient feature selection model for handwriting symbol recognition by using an improved sequential floating search method coupled with a hybrid classifier, which is obtained by combining hidden Markov models with multilayer forward network. The effectiveness of proposed method is verified by comprehensive experiments based on UNIPEN database","PeriodicalId":297071,"journal":{"name":"2006 5th International Conference on Machine Learning and Applications (ICMLA'06)","volume":"226 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134619074","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：