Browsing by Author "Thitipong Tanprasert"
Results Per Page
Sort Options
-
Item
-
ItemAn Adaptive Learning Algorithm for Supervised Neural Network with Contour Preserving Classification( 2009-11) Piyabute Fuangkhon ; Thitipong TanprasertA study of noise tolerance characteristics of an adaptive learning algorithm for supervised neural network is presented in this paper. The algorithm allows the existing knowledge to age out in slow rate as a supervised neural network is gradually retrained with consecutive sets of new samples, resembling the change of application locality under a consistent environment. The algorithm utilizes the contour preserving classification algorithm to pre-process the training data to improve the classification and the noise tolerance. The experimental results convincingly confirm the effectiveness of the algorithm and the improvement of noise tolerance.
-
Item
-
ItemClustering Analysis on Alumni Data Using Abandoned and Reborn Particle Swarm Optimization( 2016-02) Paulus Mudjihartono ; Thitipong Tanprasert ; Rachsuda JiamthapthaksinAlumni data is one of the most important data that university management uses for developing the learning process decisions. This paper applies the idea of Abandoned and Reborn PSO (AR-PSO) to convert a clustering problem into the optimization form with an objective function to minimize the ugliness of the desired clusters. This algorithm of Clustering using AR-PSO (CAR-PSO) is slightly adapted to the cluster problem domain. The generated clusters need to be examined to decide if they are acceptable. There are three evaluations; the closeness, the separation and the purity. Finally, the experiment results show that the CAR-PSO is comparable with &-means in both types of alumni data while leaving the other two clustering algorithms.
-
ItemCreatinine prediction from body composition a neural network approach( 2011) Thitipong Tanprasert ; Chularat TanprasertCreatinine, a naturally-produced chemical compound in blood, has been commonly used as a reliable indicator of kidney function. Creatinine level is typically obtained from blood-test. In this paper, a technique for predicting the criticality of creatinine level in blood is presented. The proposed technique takes only body size and mass parameters obtained from advanced weighing scale and body scanner, allowing the prediction to be done more casually. The technique applies a multi-layered feed-forward neural network for developing the prediction model. The achieved overall prediction accuracy is in the vicinity of 88% where the average false negative rate and the average false positive rate are 22.15% and 8.26%, respectively.
-
ItemDiagnosing prostate cancer using backpropagation neural network and greedy decision procedureA novel procedure for diagnosing prostate cancer (PC) based on Back propagation Neural Network (BPNN) is proposed. Elderly men with symptoms such as urinary retention, urinary hesitancy, urinary dribbling, burning urination, hematuria, etc. are considered as primary attributes. Prostate-specific antigen (PSA) level and Gleason score are the secondary attributes. Initial dataset is generated based on the clinical database. The BPNN assigns symptom levels of a set of patients based on their primary attributes. A greedy decision procedure predicts tumor stages of patients based on their strong symptom levels and secondary attributes. The simulation shows that the proposed procedure is an effective way for diagnosing prostate tumor stages.
-
ItemDistance measurement with smartphone using acceleration model of hand movement( 2016-02) Eakawat Tantamjarik ; Thitipong TanprasertThis paper proposes a novel method to obtain the displacement of a smartphone movement. The method utilizes genetic algorithm to synthesize a mathematical model of acceleration based on behavior of a person's hand movement from the raw acceleration data. Then, double integration is performed on the synthesized acceleration model, which is significantly less affected by the noise accumulation. The raw acceleration of the hand's movement is calibrated initially using acceleration-time graph analysis and a modified version of peak detection based on moving average is used to obtain the constraints for genetic algorithm. The obtained experiment results showed that the method is very effective at determining displacement with high accuracy.
-
Item
-
ItemAn incremental learning algorithm for supervised neural network with contour preserving Classification( 2009-05) Piyabute Fuangkhon ; Thitipong TanprasertThis paper presents an alternative algorithm for integrating the existing knowledge of a supervised learning neural network with the new training data. The algorithm allows the existing knowledge to age out in slow rate as a neural network is gradually retrained with consecutive sets of new samples, resembling the change of application locality under a consistent environment. The algorithm also utilizes the contour preserving classification algorithm to increase the accuracy of classification. The experiment is performed on 2-dimension partition problem and the result convincingly confirms the effectiveness of the algorithm.
-
Item
-
Item
-
ItemMulti-class contour preserving classification( 2012-08) Piyabute Fuangkhon ; Thitipong TanprasertThe original contour preserving classification technique was proposed to improve the robustness and weight fault tolerance of a neu- ral network applied with a two-class linearly separable problem. It was recently found to be improving the level of accuracy of two-class classi- fication. This paper presents an augmentation of the original technique to improve the level of accuracy of multi-class classification by better preservation of the shape or distribution model of a multi-class problem. The test results on six real world multi-class datasets from UCI ma- chine learning repository present that the proposed technique supports multi-class data and can improve the level of accuracy of multi-class classification more effectively.
-
ItemMulti-class contour preserving classification( 2012-08) Piyabute Fuangkhon ; Thitipong TanprasertThe original contour preserving classification technique was proposed to improve the robustness and weight fault tolerance of a neu- ral network applied with a two-class linearly separable problem. It was recently found to be improving the level of accuracy of two-class classi- fication. This paper presents an augmentation of the original technique to improve the level of accuracy of multi-class classification by better preservation of the shape or distribution model of a multi-class problem. The test results on six real world multi-class datasets from UCI ma- chine learning repository present that the proposed technique supports multi-class data and can improve the level of accuracy of multi-class classification more effectively.
-
Item
-
Item
-
Item
-
ItemReduced Multi-class Contour Preserving ClassificationThis research presents the augmentation of the original contour preserving classification technique to support multi-class data and to reduce the number of synthesized vectors, called multi-class outpost vectors (MCOVs). The technique has been proven to function on both synthetic-problem data sets and real-world data sets correctly. The technique also includes three methods to reduce the number of MCOVs by using minimum vector distance selection between fundamental multi-class outpost vectors and additional multi-class outpost vectors to select only MCOVs located at the decision boundary between consecutive classes of data. The three MCOV reduction methods include the FF-AA reduction method, the FA-AF reduction method, and the FAF-AFA reduction method. An evaluation has been conducted to show the reduction capability, the contour preservation capability, and the levels of classification accuracy of the three MCOV reduction methods on both non-overlapping and highly overlapping synthetic-problem data sets and highly overlapping real-world data sets. For non-overlapping problems, the experimental results present that the FA-AF reduction method can partially reduce the number of MCOVs while preserving the contour of the problem most accurately and obtaining similar levels of classification accuracy as when the whole set of MCOVs is used. For highly overlapping problems, the experimental results present that the FF-AA reduction method can partially reduce the number of MCOVs while preserving the contour of the problem most accurately and obtaining similar levels of classification accuracy as when the whole set of MCOVs is used.
-
Item
-
Item