Krung Sinapiromsaran | Chulalongkorn University (original) (raw)
Papers by Krung Sinapiromsaran
Pattern Analysis and Applications, 2016
Proceedings in Adaptation, Learning and Optimization, 2015
Proceedings in Adaptation, Learning and Optimization, 2015
International Journal of Data Mining and Bioinformatics, 2015
Class imbalance learning has recently drawn considerable attention among researchers. In this are... more Class imbalance learning has recently drawn considerable attention among researchers. In this area, a rare class is the class of primary interest from the aim of classification. Unfortunately, traditional machine learning algorithms fail to detect this class because a huge majority class overwhelms a tiny minority class. In this paper, we propose a new technique called CORE to handle the class imbalance problem. The objective of CORE is to strengthen the core of a minority class and weaken the risk of misclassified minority instances nearby the borderline of a majority class. These core and borderline regions are defined by the applicability of a safe level. As a result, a minority class is more crowed and dominant. The experiment shows that CORE can significantly improve the predictive performance of a minority class when its dataset is imbalance.
2000 Winter Simulation Conference Proceedings (Cat. No.00CH37165), 2000
We propose solving continuous parametric simulation optimizations using a deterministic nonlinear... more We propose solving continuous parametric simulation optimizations using a deterministic nonlinear optimization algorithm and sample-path simulations. The optimization problem is written in a modeling language with a simulation module accessed with an external function call. Since we allow no changes to the simulation code at all, we propose using a quadratic approximation of the simulation function to obtain derivatives. Results on three different queueing models are presented that show our method to be effective on a variety of practical problems. ©
Journal of Physics: Conference Series, 2014
The main topic of this thesis is solving simulation optimizations using a deterministicnonlinear ... more The main topic of this thesis is solving simulation optimizations using a deterministicnonlinear solver based on the sample-path optimization concept. The simulation functionis considered as a black box that deterministically returns the exact output for the sameinput value. The gradient-based nonlinear solver finds a local optimal based on thefunction and gradient evaluation of the sample path simulation function. The simulationoutput
Journal of Physics: Conference Series, 2014
The simplex algorithm is a popular algorithm for solving linear programming problems. If the orig... more The simplex algorithm is a popular algorithm for solving linear programming problems. If the origin point satisfies all constraints then the simplex can be started. Otherwise, artificial variables will be introduced to start the simplex algorithm. If we can start the simplex algorithm without using artificial variables then the simplex iterate will require less time. In this paper, we present the artificial-free technique for the simplex algorithm by mapping the problem into the objective plane and splitting constraints into three groups. In the objective plane, one of variables which has a nonzero coefficient of the objective function is fixed in terms of another variable. Then it can split constraints into three groups: the positive coefficient group, the negative coefficient group and the zero coefficient group. Along the objective direction, some constraints from the positive coefficient group will form the optimal solution. If the positive coefficient group is nonempty, the algorithm starts with relaxing constraints from the negative coefficient group and the zero coefficient group. We guarantee the feasible region obtained from the positive coefficient group to be nonempty. The transformed problem is solved using the simplex algorithm. Additional constraints from the negative coefficient group and the zero coefficient group will be added to the solved problem and use the dual simplex method to determine the new optimal solution. An example shows the effectiveness of our algorithm.
Internal sorting is a problem of finding the permutation from a list of numbers so that the appli... more Internal sorting is a problem of finding the permutation from a list of numbers so that the applied permutation list is sorted. Many sorting algorithms make use of various techniques to accomplish the sorting task. Moreover distinct characteristics of a finite list are extensively studied to find the practical sorting algorithm. This paper exhibits the unknown characteristic called the successive
We solve a simulation optimization using a deterministic nonlinear solver based on thesample-path... more We solve a simulation optimization using a deterministic nonlinear solver based on thesample-path concept. The method used a quadratic model built from a collection of surroundingsimulation points. The scheme does not require the modification of the original simulation sourcecode and is carried out automatically. Due to the large number of simulation runs, the highthroughputcomputing environment, Condor, is used. Simulation computations
Lecture Notes in Economics and Mathematical Systems, 2000
. We consider a primal-dual approach to solve nonlinear programmingproblems within the AMPL model... more . We consider a primal-dual approach to solve nonlinear programmingproblems within the AMPL modeling language, via a mixed complementarity formulation.The modeling language supplies the first order and second order derivativeinformation of the Lagrangian function of the nonlinear problem using automaticdifferentiation. The PATH solver finds the solution of the first order conditionswhich are generated automatically from this derivative information. In addition,the link incorporates the...
Computer Aided Architectural Design Futures 2005, 2005
For many decades, solving the optimal architectural layout design is unattainable for the reasona... more For many decades, solving the optimal architectural layout design is unattainable for the reasonable problem sizes. Architects have to settle for acceptable layouts instead of the favourable optimal solution. With today technologies, various optimization techniques have been used to alleviate the optimal search according to diversified goals. This paper formulates the optimal architectural layout design as the multiobjective mixed integer
Applied Mathematics and Computation, 2014
Solving a general linear programming problem using the simplex algorithm relies on introducing ar... more Solving a general linear programming problem using the simplex algorithm relies on introducing artificial variables that creates a large search space. This paper presents the non-acute constraint relaxation technique that not only eliminates the need for artificial variables but also reduces the start-up time to solve the initial relaxation problem. To guarantee the optimal solution or infeasibility or unboundedness of a linear programming problem, the algorithm reinserts the non-acute constraints back to the relaxation problem. The results of this algorithm are superior than the original simplex algorithm with artificial variables for a linear programming problem with a large number of acute constraints.
2010 International Conference on Financial Theory and Engineering, 2010
Entropy multi-hyperplane credit scoring model is a decision model that classifies applicants into... more Entropy multi-hyperplane credit scoring model is a decision model that classifies applicants into payers or defaulters by optimizing the classification cost using multiple hyperplanes based on entropy order. In the first stage, the model uses a pair of hyperplanes composed of half of the attributes which are ordered increasingly by the entropy. The hyperplanes divide the applicants into 3 groups,
2013 International Computer Science and Engineering Conference (ICSEC), 2013
Proceedings of the Annual International Conference on Computer Science Education: Innovation & Technology CSEIT 2010 & Proceedings of the Annual International Conference on Software Engineering SE 2010, Dec 6, 2010
SMOTE is an over-sampling technique for handling a class imbalanced problem. It improves the prec... more SMOTE is an over-sampling technique for handling a class imbalanced problem. It improves the precision measure of the minority class prediction by generating more minority class instances near the existing ones. Nevertheless, the large number of synthesized minority class instances may outweigh majority class instances. In this paper, we introduce the mixture techniques of over-sampling by SMOTE and under-sampling by reduction around centroids. Our algorithm, Synthetic Minority Over-Sampling and Under-sampling TEchnique called SMOUTE, avoids synthesizing a large number of minority class instances while balances both class instances. We perform experiments based on three classifiers, C4.5, Naïve Bayes and multilayer perceptron. Our results show that classifiers using SMOUTE are correctly grouped the minority class better than SMOTE. Moreover, the speed of SMOUTE is much faster than that of SMOTE for large datasets.
2012 Tenth International Conference on ICT and Knowledge Engineering, 2012
Lecture Notes in Computer Science, Apr 27, 2009
The class imbalanced problem occurs in various disciplines when one of target classes has a tiny ... more The class imbalanced problem occurs in various disciplines when one of target classes has a tiny number of instances comparing to other classes. A typical classifier normally ignores or neglects to detect a minority class due to the small number of class instances. SMOTE is one of over-sampling techniques that remedies this situation. It generates minority instances within the overlapping regions. However, SMOTE randomly synthesizes the minority instances along a line joining a minority instance and its selected nearest neighbours, ignoring nearby majority instances. Our technique called Safe-Level-SMOTE carefully samples minority instances along the same line with different weight degree, called safe level. The safe level computes by using nearest neighbour minority instances. By synthesizing the minority instances more around larger safe level, we achieve a better accuracy performance than SMOTE and Borderline-SMOTE.
2010 2nd International Conference on Computer Engineering and Technology, 2010
In this paper, we concentrate on optimizing the yearly scheduling of maintenance and inspection o... more In this paper, we concentrate on optimizing the yearly scheduling of maintenance and inspection on offshore platform for a major Petroleum company in Thailand. Efficient equipment and offshore personnel are the most significant factors in providing more efficient products and services. The optimal resource allocation problems at Quarter Platform (QP) and living barge with qualified personnel who work on specific
Pattern Analysis and Applications, 2016
Proceedings in Adaptation, Learning and Optimization, 2015
Proceedings in Adaptation, Learning and Optimization, 2015
International Journal of Data Mining and Bioinformatics, 2015
Class imbalance learning has recently drawn considerable attention among researchers. In this are... more Class imbalance learning has recently drawn considerable attention among researchers. In this area, a rare class is the class of primary interest from the aim of classification. Unfortunately, traditional machine learning algorithms fail to detect this class because a huge majority class overwhelms a tiny minority class. In this paper, we propose a new technique called CORE to handle the class imbalance problem. The objective of CORE is to strengthen the core of a minority class and weaken the risk of misclassified minority instances nearby the borderline of a majority class. These core and borderline regions are defined by the applicability of a safe level. As a result, a minority class is more crowed and dominant. The experiment shows that CORE can significantly improve the predictive performance of a minority class when its dataset is imbalance.
2000 Winter Simulation Conference Proceedings (Cat. No.00CH37165), 2000
We propose solving continuous parametric simulation optimizations using a deterministic nonlinear... more We propose solving continuous parametric simulation optimizations using a deterministic nonlinear optimization algorithm and sample-path simulations. The optimization problem is written in a modeling language with a simulation module accessed with an external function call. Since we allow no changes to the simulation code at all, we propose using a quadratic approximation of the simulation function to obtain derivatives. Results on three different queueing models are presented that show our method to be effective on a variety of practical problems. ©
Journal of Physics: Conference Series, 2014
The main topic of this thesis is solving simulation optimizations using a deterministicnonlinear ... more The main topic of this thesis is solving simulation optimizations using a deterministicnonlinear solver based on the sample-path optimization concept. The simulation functionis considered as a black box that deterministically returns the exact output for the sameinput value. The gradient-based nonlinear solver finds a local optimal based on thefunction and gradient evaluation of the sample path simulation function. The simulationoutput
Journal of Physics: Conference Series, 2014
The simplex algorithm is a popular algorithm for solving linear programming problems. If the orig... more The simplex algorithm is a popular algorithm for solving linear programming problems. If the origin point satisfies all constraints then the simplex can be started. Otherwise, artificial variables will be introduced to start the simplex algorithm. If we can start the simplex algorithm without using artificial variables then the simplex iterate will require less time. In this paper, we present the artificial-free technique for the simplex algorithm by mapping the problem into the objective plane and splitting constraints into three groups. In the objective plane, one of variables which has a nonzero coefficient of the objective function is fixed in terms of another variable. Then it can split constraints into three groups: the positive coefficient group, the negative coefficient group and the zero coefficient group. Along the objective direction, some constraints from the positive coefficient group will form the optimal solution. If the positive coefficient group is nonempty, the algorithm starts with relaxing constraints from the negative coefficient group and the zero coefficient group. We guarantee the feasible region obtained from the positive coefficient group to be nonempty. The transformed problem is solved using the simplex algorithm. Additional constraints from the negative coefficient group and the zero coefficient group will be added to the solved problem and use the dual simplex method to determine the new optimal solution. An example shows the effectiveness of our algorithm.
Internal sorting is a problem of finding the permutation from a list of numbers so that the appli... more Internal sorting is a problem of finding the permutation from a list of numbers so that the applied permutation list is sorted. Many sorting algorithms make use of various techniques to accomplish the sorting task. Moreover distinct characteristics of a finite list are extensively studied to find the practical sorting algorithm. This paper exhibits the unknown characteristic called the successive
We solve a simulation optimization using a deterministic nonlinear solver based on thesample-path... more We solve a simulation optimization using a deterministic nonlinear solver based on thesample-path concept. The method used a quadratic model built from a collection of surroundingsimulation points. The scheme does not require the modification of the original simulation sourcecode and is carried out automatically. Due to the large number of simulation runs, the highthroughputcomputing environment, Condor, is used. Simulation computations
Lecture Notes in Economics and Mathematical Systems, 2000
. We consider a primal-dual approach to solve nonlinear programmingproblems within the AMPL model... more . We consider a primal-dual approach to solve nonlinear programmingproblems within the AMPL modeling language, via a mixed complementarity formulation.The modeling language supplies the first order and second order derivativeinformation of the Lagrangian function of the nonlinear problem using automaticdifferentiation. The PATH solver finds the solution of the first order conditionswhich are generated automatically from this derivative information. In addition,the link incorporates the...
Computer Aided Architectural Design Futures 2005, 2005
For many decades, solving the optimal architectural layout design is unattainable for the reasona... more For many decades, solving the optimal architectural layout design is unattainable for the reasonable problem sizes. Architects have to settle for acceptable layouts instead of the favourable optimal solution. With today technologies, various optimization techniques have been used to alleviate the optimal search according to diversified goals. This paper formulates the optimal architectural layout design as the multiobjective mixed integer
Applied Mathematics and Computation, 2014
Solving a general linear programming problem using the simplex algorithm relies on introducing ar... more Solving a general linear programming problem using the simplex algorithm relies on introducing artificial variables that creates a large search space. This paper presents the non-acute constraint relaxation technique that not only eliminates the need for artificial variables but also reduces the start-up time to solve the initial relaxation problem. To guarantee the optimal solution or infeasibility or unboundedness of a linear programming problem, the algorithm reinserts the non-acute constraints back to the relaxation problem. The results of this algorithm are superior than the original simplex algorithm with artificial variables for a linear programming problem with a large number of acute constraints.
2010 International Conference on Financial Theory and Engineering, 2010
Entropy multi-hyperplane credit scoring model is a decision model that classifies applicants into... more Entropy multi-hyperplane credit scoring model is a decision model that classifies applicants into payers or defaulters by optimizing the classification cost using multiple hyperplanes based on entropy order. In the first stage, the model uses a pair of hyperplanes composed of half of the attributes which are ordered increasingly by the entropy. The hyperplanes divide the applicants into 3 groups,
2013 International Computer Science and Engineering Conference (ICSEC), 2013
Proceedings of the Annual International Conference on Computer Science Education: Innovation & Technology CSEIT 2010 & Proceedings of the Annual International Conference on Software Engineering SE 2010, Dec 6, 2010
SMOTE is an over-sampling technique for handling a class imbalanced problem. It improves the prec... more SMOTE is an over-sampling technique for handling a class imbalanced problem. It improves the precision measure of the minority class prediction by generating more minority class instances near the existing ones. Nevertheless, the large number of synthesized minority class instances may outweigh majority class instances. In this paper, we introduce the mixture techniques of over-sampling by SMOTE and under-sampling by reduction around centroids. Our algorithm, Synthetic Minority Over-Sampling and Under-sampling TEchnique called SMOUTE, avoids synthesizing a large number of minority class instances while balances both class instances. We perform experiments based on three classifiers, C4.5, Naïve Bayes and multilayer perceptron. Our results show that classifiers using SMOUTE are correctly grouped the minority class better than SMOTE. Moreover, the speed of SMOUTE is much faster than that of SMOTE for large datasets.
2012 Tenth International Conference on ICT and Knowledge Engineering, 2012
Lecture Notes in Computer Science, Apr 27, 2009
The class imbalanced problem occurs in various disciplines when one of target classes has a tiny ... more The class imbalanced problem occurs in various disciplines when one of target classes has a tiny number of instances comparing to other classes. A typical classifier normally ignores or neglects to detect a minority class due to the small number of class instances. SMOTE is one of over-sampling techniques that remedies this situation. It generates minority instances within the overlapping regions. However, SMOTE randomly synthesizes the minority instances along a line joining a minority instance and its selected nearest neighbours, ignoring nearby majority instances. Our technique called Safe-Level-SMOTE carefully samples minority instances along the same line with different weight degree, called safe level. The safe level computes by using nearest neighbour minority instances. By synthesizing the minority instances more around larger safe level, we achieve a better accuracy performance than SMOTE and Borderline-SMOTE.
2010 2nd International Conference on Computer Engineering and Technology, 2010
In this paper, we concentrate on optimizing the yearly scheduling of maintenance and inspection o... more In this paper, we concentrate on optimizing the yearly scheduling of maintenance and inspection on offshore platform for a major Petroleum company in Thailand. Efficient equipment and offshore personnel are the most significant factors in providing more efficient products and services. The optimal resource allocation problems at Quarter Platform (QP) and living barge with qualified personnel who work on specific