Pub Date : 2019-10-01DOI: 10.1109/BRACIS.2019.00058
Edson Lopes da Silva Junior, V. S. Silva, L. B. Gonçalves, L. Moreno, S. Soares
This paper addresses the Tabu Clustered Traveling Salesman Problem, a recent variation of TSP motivated by a real-life application in the management of satellite network. In the TCTSP, nodes are split into two distinct partitions, named tabus and clusters. The TCTSP consists of finding a minimum cost cycle that contains exactly one node of each tabu and visits nodes of the same cluster consecutively. We propose a reactive GRASP heuristic that is competitive in terms of both time efficiency and solution quality in comparison with the literature results.
{"title":"An Efficient Algorithm for the Tabu Clustered Traveling Salesman Problem","authors":"Edson Lopes da Silva Junior, V. S. Silva, L. B. Gonçalves, L. Moreno, S. Soares","doi":"10.1109/BRACIS.2019.00058","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00058","url":null,"abstract":"This paper addresses the Tabu Clustered Traveling Salesman Problem, a recent variation of TSP motivated by a real-life application in the management of satellite network. In the TCTSP, nodes are split into two distinct partitions, named tabus and clusters. The TCTSP consists of finding a minimum cost cycle that contains exactly one node of each tabu and visits nodes of the same cluster consecutively. We propose a reactive GRASP heuristic that is competitive in terms of both time efficiency and solution quality in comparison with the literature results.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116296173","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00057
J. Carvalho, P. Farias, E. Souza, E. F. S. Filho
Mobile robot localization is a complex task, specially in unstructured indoor environments, due to noise and wrong data association from sensors. The localization procedure is even harder when the vehicle has low confidence about its last pose estimate, situation that requires a Global Localization procedure. In this work, a Global Localization algorithm based on Particle Swarm Optimization (PSO) is integrated with a Pose Tracking algorithm, the Perfect Match, to obtain a robust localization technique. Results show that this approach can solve the problem of Global Localization with good performance.
{"title":"Particle Swarm Localization for Mobile Robots Using a 2D Laser Sensor","authors":"J. Carvalho, P. Farias, E. Souza, E. F. S. Filho","doi":"10.1109/BRACIS.2019.00057","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00057","url":null,"abstract":"Mobile robot localization is a complex task, specially in unstructured indoor environments, due to noise and wrong data association from sensors. The localization procedure is even harder when the vehicle has low confidence about its last pose estimate, situation that requires a Global Localization procedure. In this work, a Global Localization algorithm based on Particle Swarm Optimization (PSO) is integrated with a Pose Tracking algorithm, the Perfect Match, to obtain a robust localization technique. Results show that this approach can solve the problem of Global Localization with good performance.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131855245","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00038
V. B. D. Santos, L. N. Barros, Maria Viviane de Menezes
A solution for Fully Observable Non-Deterministic FOND planning is classified as weak, strong or strong-cyclic, indicating different ways to reach a goal state due to the non-determinism of the actions. In this paper, we propose a new algorithm to solve FOND planning problems for strong-cyclic policies, where the agent always achieves the goal, under the fairness assumption that execution will eventually exit from all existing cycles. The planner, called Pactl-Sym-StrongCyclic, is based on symbolic model checking using alpha-CTL logic: an extension of CTL that considers actions behind the transitions. To the best of our knowledge, this is the first FOND planner for strong-cyclic policies that applies symbolic reasoning over the actions (and not over the state transition relation), and therefore can outperform state-of-the-art FOND planners.
完全可观察非确定性FOND规划的解决方案分为弱、强或强循环，表明由于行动的不确定性而达到目标状态的不同方法。本文提出了一种新的算法来解决强循环策略中agent总能达到目标的FOND规划问题，该算法在公平性假设下执行最终会退出所有现有循环。称为pactl - sm - strongcyclic的计划器基于使用alpha-CTL逻辑的符号模型检查:CTL的扩展，考虑转换背后的操作。据我们所知，这是第一个对动作(而不是状态转换关系)应用符号推理的强循环策略的FOND规划器，因此可以优于最先进的FOND规划器。
{"title":"Symbolic Planning for Strong-Cyclic Policies","authors":"V. B. D. Santos, L. N. Barros, Maria Viviane de Menezes","doi":"10.1109/BRACIS.2019.00038","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00038","url":null,"abstract":"A solution for Fully Observable Non-Deterministic FOND planning is classified as weak, strong or strong-cyclic, indicating different ways to reach a goal state due to the non-determinism of the actions. In this paper, we propose a new algorithm to solve FOND planning problems for strong-cyclic policies, where the agent always achieves the goal, under the fairness assumption that execution will eventually exit from all existing cycles. The planner, called Pactl-Sym-StrongCyclic, is based on symbolic model checking using alpha-CTL logic: an extension of CTL that considers actions behind the transitions. To the best of our knowledge, this is the first FOND planner for strong-cyclic policies that applies symbolic reasoning over the actions (and not over the state transition relation), and therefore can outperform state-of-the-art FOND planners.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"141 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122852210","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00091
Lucas Rodrigues Carneiro, Carla Delgado, J. F. C. Silva
Video games normally use Artificial Intelligence techniques to improve Non-Player Character (NPC) behavior, creating a more realistic experience for their players. However, rational behavior in general does not consider social interactions between player and bots. Because of that, a new framework for NPCs was proposed, which uses a social bias to mix the default strategy of finding the best possible plays to win with a analysis to decide if other players should be categorized as allies or foes. Trust and reputation models were used together to implement this demeanor. In this paper we discuss an implementation of this framework inside the game Settlers of Catan. New NPC agents are created to this implementation. We also analyze the results obtained from simulations among agents and players to conclude how the use of trust and reputation in NPCs can create a better gaming experience.
{"title":"Social Analysis of Game Agents: How Trust and Reputation can Improve Player Experience","authors":"Lucas Rodrigues Carneiro, Carla Delgado, J. F. C. Silva","doi":"10.1109/BRACIS.2019.00091","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00091","url":null,"abstract":"Video games normally use Artificial Intelligence techniques to improve Non-Player Character (NPC) behavior, creating a more realistic experience for their players. However, rational behavior in general does not consider social interactions between player and bots. Because of that, a new framework for NPCs was proposed, which uses a social bias to mix the default strategy of finding the best possible plays to win with a analysis to decide if other players should be categorized as allies or foes. Trust and reputation models were used together to implement this demeanor. In this paper we discuss an implementation of this framework inside the game Settlers of Catan. New NPC agents are created to this implementation. We also analyze the results obtained from simulations among agents and players to conclude how the use of trust and reputation in NPCs can create a better gaming experience.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122964406","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00059
M. A. A. Kappel
Action scheduling optimization is a problem that involves chronologically organizing a set of actions, jobs or commands in order to accomplish a pre-established goal. This kind of problem can be found in a number of areas, such as production planning, delivery logistic organization, robot movement planning and behavior programming for intelligent agents in games. Despite being a recurrent problem, selecting the appropriate time and order to execute each task is not trivial, and typically involves highly complex techniques. The main objective of this work is to provide a simple alternative to tackle the action scheduling problem, by using Cartesian Genetic Programming as an approach. The proposed solution involves the application of two simple main steps: defining the set of available actions and specifying an objective function to be optimized. Then, by the means of the evolutionary algorithm, an automatically generated schedule will be revealed as the most fitting to the goal. The effectiveness of this methodology was tested by performing an action schedule optimization on two different problems involving virtual agents walking in a simulated environment. In both cases, results showed that, throughout the evolutionary process, the simulated agents naturally chose the most efficient sequential and parallel combination of actions to reach greater distances. The use of evolutionary adaptive metaheuristics such as Cartesian Genetic Programming allows the identification of the best possible schedule of actions to solve a problem.
{"title":"Action Scheduling Optimization using Cartesian Genetic Programming","authors":"M. A. A. Kappel","doi":"10.1109/BRACIS.2019.00059","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00059","url":null,"abstract":"Action scheduling optimization is a problem that involves chronologically organizing a set of actions, jobs or commands in order to accomplish a pre-established goal. This kind of problem can be found in a number of areas, such as production planning, delivery logistic organization, robot movement planning and behavior programming for intelligent agents in games. Despite being a recurrent problem, selecting the appropriate time and order to execute each task is not trivial, and typically involves highly complex techniques. The main objective of this work is to provide a simple alternative to tackle the action scheduling problem, by using Cartesian Genetic Programming as an approach. The proposed solution involves the application of two simple main steps: defining the set of available actions and specifying an objective function to be optimized. Then, by the means of the evolutionary algorithm, an automatically generated schedule will be revealed as the most fitting to the goal. The effectiveness of this methodology was tested by performing an action schedule optimization on two different problems involving virtual agents walking in a simulated environment. In both cases, results showed that, throughout the evolutionary process, the simulated agents naturally chose the most efficient sequential and parallel combination of actions to reach greater distances. The use of evolutionary adaptive metaheuristics such as Cartesian Genetic Programming allows the identification of the best possible schedule of actions to solve a problem.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"101 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121849152","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00154
Mateus Teixeira Magalhães, G. Miranda, L. B. Gonçalves, L. Moreno, S. Soares, L. W. Oliveira
Power distribution companies around the world have been looking for ways to improve the quality of power transportation through their distribution networks, thereby improving the quality of product delivery to their consumers. The growing automation of distribution systems brings the possibility of changing the network configuration more easily, by means of maneuvers of the sectioning devices, enabling actions that allow the system to operate in the most adequate way, with reduction in losses and improvement in loading and voltage levels. The study of a configuration that minimizes losses by Joule effect has extreme importance to improve the performance of a system, with lower cost and still more sustainability avoiding waste. This work proposes the use of a random-key genetic algorithm implemented in a low-level language to obtain a good reconfiguration of the distribution system. The results are presented and compared with other works from literature, for different systems, aiming at obtaining a lower computational cost.
{"title":"A Random Keys Genetic Algorithm Approach for Reconfiguration Problem in Distribution Power Networks","authors":"Mateus Teixeira Magalhães, G. Miranda, L. B. Gonçalves, L. Moreno, S. Soares, L. W. Oliveira","doi":"10.1109/BRACIS.2019.00154","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00154","url":null,"abstract":"Power distribution companies around the world have been looking for ways to improve the quality of power transportation through their distribution networks, thereby improving the quality of product delivery to their consumers. The growing automation of distribution systems brings the possibility of changing the network configuration more easily, by means of maneuvers of the sectioning devices, enabling actions that allow the system to operate in the most adequate way, with reduction in losses and improvement in loading and voltage levels. The study of a configuration that minimizes losses by Joule effect has extreme importance to improve the performance of a system, with lower cost and still more sustainability avoiding waste. This work proposes the use of a random-key genetic algorithm implemented in a low-level language to obtain a good reconfiguration of the distribution system. The results are presented and compared with other works from literature, for different systems, aiming at obtaining a lower computational cost.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129990392","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00033
Caio Luiggy Riyoichi Sawada Ueno, Diego Furtado Silva
Automatic music organization and retrieval is a highly required task nowadays. Labeling songs with summarized but descriptive information have implications in a wide range of tasks in this scenario. The genre is one of the most common labels used for music recordings. Using this piece of information, music platforms can organize collections by, for instance, associating songs and artists with similar characteristics. Lyrics represent an alternative source of data for genre recognition. While "traditional" bag-of-words-based text mining techniques represent a considerable part of the literature, recent papers shown an advance on this task applying deep learning algorithms. However, there is no research on how these distinct strategies contribute to each other. In this paper, we explore different strategies for music genre classification from lyrics and show that even simple combinations of these strategies allow improving accuracy on the lyrics-based music genre identification.
{"title":"On Combining Diverse Models for Lyrics-Based Music Genre Classification","authors":"Caio Luiggy Riyoichi Sawada Ueno, Diego Furtado Silva","doi":"10.1109/BRACIS.2019.00033","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00033","url":null,"abstract":"Automatic music organization and retrieval is a highly required task nowadays. Labeling songs with summarized but descriptive information have implications in a wide range of tasks in this scenario. The genre is one of the most common labels used for music recordings. Using this piece of information, music platforms can organize collections by, for instance, associating songs and artists with similar characteristics. Lyrics represent an alternative source of data for genre recognition. While \"traditional\" bag-of-words-based text mining techniques represent a considerable part of the literature, recent papers shown an advance on this task applying deep learning algorithms. However, there is no research on how these distinct strategies contribute to each other. In this paper, we explore different strategies for music genre classification from lyrics and show that even simple combinations of these strategies allow improving accuracy on the lyrics-based music genre identification.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"2014 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127511308","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00051
F. M. Junior, T. P. D. Araujo, J. V. M. Sousa, N. J. C. D. Costa, R. Melo, A. Pinto, A. Saraiva
This work introduces a method for recognizing handwritten polynomials using Convolutional Neural Networks (CNN) and Fractional Order Darwinian Particle Swarm Optimization (FODPSO). Segmentation of the input image is done with the FODPSO technique, which uses fractional derivative to control the rate of particle convergence. After segmentation, three CNN are used in the character recognition step: the first one classifies the individual symbols as numeric or non-numeric. The second network recognizes the numbers, while the third CNN recognize the non-numeric symbols. A heuristic procedure is used to build the polynomial, whose graph is finally plotted. A total of 264780 images containing symbols and numbers were used for training, validating, and testing the CNN, with an accuracy of approximately 99%.
{"title":"Recognition of Simple Handwritten Polynomials Using Segmentation with Fractional Calculus and Convolutional Neural Networks","authors":"F. M. Junior, T. P. D. Araujo, J. V. M. Sousa, N. J. C. D. Costa, R. Melo, A. Pinto, A. Saraiva","doi":"10.1109/BRACIS.2019.00051","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00051","url":null,"abstract":"This work introduces a method for recognizing handwritten polynomials using Convolutional Neural Networks (CNN) and Fractional Order Darwinian Particle Swarm Optimization (FODPSO). Segmentation of the input image is done with the FODPSO technique, which uses fractional derivative to control the rate of particle convergence. After segmentation, three CNN are used in the character recognition step: the first one classifies the individual symbols as numeric or non-numeric. The second network recognizes the numbers, while the third CNN recognize the non-numeric symbols. A heuristic procedure is used to build the polynomial, whose graph is finally plotted. A total of 264780 images containing symbols and numbers were used for training, validating, and testing the CNN, with an accuracy of approximately 99%.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127583759","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00075
Gabriel Mariano de Castro SIlva, Jaime Simão Sichman
In many domains, applications use people trajectories' spatiotemporal data for impersonation fraud detection purposes. Anomaly-based approaches consist in constructing mobility profiles based on users' frequent paths and schedules and comparing new trajectories against these profiles: if some new trajectory is not consistent with the user profile, a potential impersonation is detected. Previous studies, however, do not include traveling companions in users' profiles, although performing activities in social groups is inherent to human behavior. Physical access control systems on smart buildings can provide activity companions information since social groups naturally emerge on organizations hosted in such buildings and these systems can capture group trajectories. This paper explores the feasibility of using spatiotemporal mobility profiles enriched with group trajectory pattern data as a novel framework for impersonation fraud detection in smart buildings. An empirical analysis is conducted, and results show that it is feasible to add companions activities information to mobility profiles in order to enhance anomaly-based impersonation attack detection.
{"title":"Using Social Group Trajectories for Potential Impersonation Detection on Smart Buildings Access Control","authors":"Gabriel Mariano de Castro SIlva, Jaime Simão Sichman","doi":"10.1109/BRACIS.2019.00075","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00075","url":null,"abstract":"In many domains, applications use people trajectories' spatiotemporal data for impersonation fraud detection purposes. Anomaly-based approaches consist in constructing mobility profiles based on users' frequent paths and schedules and comparing new trajectories against these profiles: if some new trajectory is not consistent with the user profile, a potential impersonation is detected. Previous studies, however, do not include traveling companions in users' profiles, although performing activities in social groups is inherent to human behavior. Physical access control systems on smart buildings can provide activity companions information since social groups naturally emerge on organizations hosted in such buildings and these systems can capture group trajectories. This paper explores the feasibility of using spatiotemporal mobility profiles enriched with group trajectory pattern data as a novel framework for impersonation fraud detection in smart buildings. An empirical analysis is conducted, and results show that it is feasible to add companions activities information to mobility profiles in order to enhance anomaly-based impersonation attack detection.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"16 5","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114060328","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 : 2019-10-01DOI: 10.1109/BRACIS.2019.00042
Paulo Vitor Freitas da Silva, R. F. Neto, Saulo Moraes Villela
Subset selection is an important task in many problems, especially when dealing with high dimensional problems, such as classification, regression, and others. In this sense, this work proposes an ordered search to select variables in orthogonal regression problems based on support vectors. The admissible search is based on a monotone property of the radius parameter. Thus, we use the radius of the SV-regression as an evaluation measure for the search, making it able to find the subsets with the smallest radius in each dimension of the problem without exhaustively exploring all possibilities. The main reason for choosing the orthogonal regression is due to the fact that this model also considers the existence of error in dependent variables. The obtained results, represented by the test error, when compared to the LASSO and a recursive feature elimination technique, demonstrate the efficiency of the method.
{"title":"An Ordered Search for Subset Selection in Support Vector Orthogonal Regression","authors":"Paulo Vitor Freitas da Silva, R. F. Neto, Saulo Moraes Villela","doi":"10.1109/BRACIS.2019.00042","DOIUrl":"https://doi.org/10.1109/BRACIS.2019.00042","url":null,"abstract":"Subset selection is an important task in many problems, especially when dealing with high dimensional problems, such as classification, regression, and others. In this sense, this work proposes an ordered search to select variables in orthogonal regression problems based on support vectors. The admissible search is based on a monotone property of the radius parameter. Thus, we use the radius of the SV-regression as an evaluation measure for the search, making it able to find the subsets with the smallest radius in each dimension of the problem without exhaustively exploring all possibilities. The main reason for choosing the orthogonal regression is due to the fact that this model also considers the existence of error in dependent variables. The obtained results, represented by the test error, when compared to the LASSO and a recursive feature elimination technique, demonstrate the efficiency of the method.","PeriodicalId":335206,"journal":{"name":"Brazilian Conference on Intelligent Systems","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116790240","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：