Jin-Il Park

TAIFEX and KOSPI 200 forecasting based on two-factors high-order fuzzy time series and particle swarm optimization

Since the fuzzy time series forecasting methods provide a powerful framework to cope with vague or ambiguous problems, they have been widely used in real applications. The forecasting accuracy of these methods usually, however, depend on their universe of discourse and the length of intervals. So, we present a new forecasting method using two-factors high-order fuzzy time series and particle swarm optimization (PSO) for increasing the forecasting accuracy. To show the effectiveness of the proposed method, we applied our method for the Taiwan futures exchange (TAIFEX) forecasting and the Korea composite price index (KOSPI) 200 forecasting. The results show better forecasting accuracy than previous methods.

Hybrid Feature Selection Using Genetic Algorithm and Information Theory

In pattern classification, feature selection is an important factor in the performance of classifiers. In particular, when classifying a large number of features or variables, the accuracy and computational time of the classifier can be improved by using the relevant feature subset to remove the irrelevant, redundant, or noisy data. The proposed method consists of two parts: a wrapper part with an improved genetic algorithm(GA) using a new reproduction method and a filter part using mutual information. We also considered feature selection methods based on mutual information(MI) to improve computational complexity. Experimental results show that this method can achieve better performance in pattern recognition problems than other conventional solutions.

Neuro-Fuzzy Rule Generation for Backing up Navigation of Car-like Mobile Robots

An automatic neuro-fuzzy rule generation scheme is proposed for backing up navigation of car-like mobile robots. The proposed method is based on the Conditional Fuzzy C-Means (CFCM) and Fuzzy Equalization (FE) methods. The CFCM is adopted to render clusters, which can represent the homogeneous properties of the given input and output fuzzy data, and also the FE method is used to systematically construct the fuzzy membership functions for the ANFIS. From these, a compact size of fuzzy rules can be automatically obtained, which satisfy the given goal. The proposed method has been applied to a truck, and also to a truck-trailer backing up navigation problem, and good results have been achieved in comparison to previous work.

Biometric and Identity Reference Protection

This paper describes how to protect the personal information of a biometric reference provider wherein biometric reference and personally identifiable information are bounded in a biometric system. To overcome the shortcomings of the simple personal authentication method using a password, such as identify theft, a biometric system that utilizes physical and behavioral characteristics of each person is usually adopted. In the biometric system, the biometric information itself is personal information, and it can be used as an unique identifier that can identify a particular individual when combining with the other information. As a result, secure protection methods are required for generating, storing, and transmitting biometric information. Considering these issues, this paper proposes a method for ensuring confidentiality and integrity in storing and transferring personally identifiable information that is used in conjunction with biometric information, by extending the related X9.84 standard. This paper also outlines the usefulness of the proposition by defining a standard format represented by ASN.1, and implementing it.

Feature Extraction and Classification of High Dimensional Biomedical Spectral Data

In this paper, we propose the biomedical spectral pattern classification techniques by the fusion scheme based on the SpPCA and MLP in extended feature space. A conventional PCA technique for the dimension reduction has the problem that it can`t find an optimal transformation matrix if the property of input data is nonlinear. To overcome this drawback, we extract features by the SpPCA technique in extended space which use the local patterns rather than whole patterns. In the classification step, individual classifier based on MLP calculates the similarity of each class for local features. Finally, biomedical spectral patterns is classified by the fusion scheme to effectively combine the individual information. As the simulation results to verify the effectiveness, the proposed method showed more improved classification results than conventional methods.

Gait Recognition and Person Identification for Surveillance Robots

The surveillance robot has been an important component in the field of service robot industry. In the surveillance robot technology, one of the most important technology is to identify a person. In this paper, we propose a gait recognition method based on contourlet and fuzzy LDA (Linear Discriminant Analysis) for surveillance robots. After decomposing a gait image into directional subband images by contourlet, features are obtained in each subband by the fuzzy LDA. The final gait recognition is performed by a fusion technique that effectively combines similarities calculated respectively in each local subband. To show the effectiveness of the proposed algorithm, various experiments are performed for CBNU and NLPR DB datasets. From these, we obtained better classification rates in comparison with the result produced by previous methods.

Generalized Cluster Based Fuzzy Model Tree for data modeling and prediction

In this paper, we propose a novel tree based modeling method, Generalized Cluster based Fuzzy Model Tree (G-CFMT) which can model piecewise linear or piecewise nonlinear dataset and predict a continuous output value. To construct the G-CFMT, data cluster centers are calculated by fuzzy clustering and Extreme Learning Machine (ELM) are obtained at the tree nodes. Since the fuzzy clustering method can render the granulation of dataset, the complexity of the constructed tree is usually low. Moreover, we show that the ELM based scheme can also produce a linear regression model. In the prediction step, fuzzy membership values are calculated from the distance between input data and all cluster centers, the passing nodes from root to the leaf node. Final data prediction is performed by fusing the intermediate induction results, which renders capability of overcoming over-fitting problem of deteriorating the performance for testing data. To validate the proposed method, we have applied our method to various real world datasets. The experimental results clearly underline better performance over other conventional linear and nonlinear modeling and prediction methods in terms of several performance indices.

Imbedded Type Real-Time Fault Diagnosis for BLDC Motors

In this paper, we propose a fault diagnosis algorithm for BLDC motors by principle component analysis (PCA) and implement a real-time fault diagnosis system for BLDC motors. To verify the proposed diagnosis algorithm, various faulty data are acquired by Lab VIEW program from experimental system. We extract a fault feature using principle component analysis after preprocessing and then finally the fault diagnosis is performed by Euclidean similarity. Also, we embed the PCA algorithm and k-NN classification algorithm into a digital signal processor. From various experiments, we found that the proposed algorithm can be used as a powerful technique to classify the several fault signals acquired from BLDC motors.

Biometric Features and Responsible Person Information Hiding by Watermarking Technique

This paper propose a method to hide not only biometric features in the biometric image such as face and fingerprint for protecting them from unauthorized entity but also information of responsible person expressed as binary image which can be used to identify the responsibility of divulgence. For this, we investigate the recognition rates and bit error rates of extracted responsible person information watermark for the cases of using face and fingerprint images as cover images for fingerprint and face recognition which are the most popular biometric techniques. From these experiments, we confirm that the proposed method can be used for various application requiring to protect personal biometric information

Stepwise Constructive Method for Neural Networks Using a Flexible Incremental Algorithm

There have been much difficulties to construct an optimized neural network in complex nonlinear regression problems such as selecting the networks structure and avoiding overtraining problem generated by noise. In this paper, we propose a stepwise constructive method for neural networks using a flexible incremental algorithm. When the hidden nodes are added, the flexible incremental algorithm adaptively controls the number of hidden nodes by a validation dataset for minimizing the prediction residual error. Here, the ELM (Extreme Learning Machine) was used for fast training. The proposed neural network can be an universal approximator without user intervene in the training process, but also it has faster training and smaller number of hidden nodes. From the experimental results with various benchmark datasets, the proposed method shows better performance for real-world regression problems than previous methods.