Kyung-Joong Kim

Ensemble classifiers based on correlation analysis for DNA microarray classification

Abstract Since accurate classification of DNA microarray is a very important issue for the treatment of cancer, it is more desirable to make a decision by combining the results of various expert classifiers rather than by depending on the result of only one classifier. In spite of the many advantages of mutually error-correlated ensemble classifiers, they are limited in performance. It is difficult to create an optimal ensemble for DNA analysis that deals with few samples with large features. Usually, different feature sets are provided to learn the components of the ensemble expecting the improvement of classifiers. If the feature sets provide similar information, the combination of the classifiers trained from them cannot improve the performance because they will make the same error and there is no possibility of compensation. In this paper, we adopt correlation analysis of feature selection methods as a guideline of the separation of features to learn the components of ensemble. We propose two different correlation methods for the generation of feature sets to learn ensemble classifiers. Each ensemble classifier combines several other classifiers learned from different features and based on correlation analysis to classify cancer precisely. In this way, it is possible to systematically evaluate the performance of the proposed method with three benchmark datasets. Experimental results show that two ensemble classifiers whose components are learned from different feature sets that are negatively or complementarily correlated with each other produce the best recognition rates on the three benchmark datasets.

A Comprehensive Overview of the Applications of Artificial Life

We review the applications of artificial life (ALife), the creation of synthetic life on computers to study, simulate, and understand living systems. The definition and features of ALife are shown by application studies. ALife application fields treated include robot control, robot manufacturing, practical robots, computer graphics, natural phenomenon modeling, entertainment, games, music, economics, Internet, information processing, industrial design, simulation software, electronics, security, data mining, and telecommunications. In order to show the status of ALife application research, this review primarily features a survey of about 180 ALife application articles rather than a selected representation of a few articles. Evolutionary computation is the most popular method for designing such applications, but recently swarm intelligence, artificial immune network, and agent-based modeling have also produced results. Applications were initially restricted to the robotics and computer graphics, but presently, many different applications in engineering areas are of interest.

AniDiary: Daily Cartoon-Style Diary Exploits Bayesian Networks

AniDiary (Anywhere Diary) uses Bayesian networks to automatically detect landmark events and summarize a users daily life in a cartoon-style diary. Our goal was to summarize a given users daily life with a cartoon-style diary based on information collected from mobile devices such as smart phones. Our system, AniDiary (Anywhere Diary), addresses two main problems of typical diary systems-the huge number of events originating from the real-life log and the awkward presentation of the output. Using modular Bayesian networks, AniDiary can detect and visualize landmarks (relevant or novel events) and transform numerous logs into user-friendly cartoon images. The cartoons provide a good starting point for fine-grained searches of detailed information. For example, users can link each cartoon to rich media (photos or videos) that offer more details, reducing the search space and letting users easily recall the linked details.

Personalized mining of web documents using link structures and fuzzy concept networks

Personalized search engines are important tools for finding web documents for specific users, because they are able to provide the location of information on the WWW as accurately as possible, using efficient methods of data mining and knowledge discovery. The types and features of traditional search engines are various, including support for different functionality and ranking methods. New search engines that use link structures have produced improved search results which can overcome the limitations of conventional text-based search engines. Going a step further, this paper presents a system that provides users with personalized results derived from a search engine that uses link structures. The fuzzy document retrieval system (constructed from a fuzzy concept network based on the users profile) personalizes the results yielded from link-based search engines with the preferences of the specific user. A preliminary experiment with six subjects indicates that the developed system is capable of searching not only relevant but also personalized web pages, depending on the preferences of the user.

An Evolutionary Algorithm Approach to Optimal Ensemble Classifiers for DNA Microarray Data Analysis

In general, the analysis of microarray data requires two steps: feature selection and classification. From a variety of feature selection methods and classifiers, it is difficult to find optimal ensembles composed of any feature-classifier pairs. This paper proposes a novel method based on the evolutionary algorithm (EA) to form sophisticated ensembles of features and classifiers that can be used to obtain high classification performance. In spite of the exponential number of possible ensembles of individual feature-classifier pairs, an EA can produce the best ensemble in a reasonable amount of time. The chromosome is encoded with real values to decide the weight for each feature-classifier pair in an ensemble. Experimental results with two well-known microarray datasets in terms of time and classification rate indicate that the proposed method produces ensembles that are superior to individual classifiers, as well as other ensembles optimized by random and greedy strategies.

Prediction of colon cancer using an evolutionary neural network

Colon cancer is second only to lung cancer as a cause of cancer-related mortality in Western countries. Colon cancer is a genetic disease, propagated by the acquisition of somatic alterations that influence gene expression. DNA microarray technology provides a format for the simultaneous measurement of the expression level of thousands of genes in a single hybridization assay. The most exciting result of microarray technology has been the demonstration that patterns of gene expression can distinguish between tumors of different anatomical origin. Standard statistical methodologies in classification and prediction do not work well or even at all when N (a number of samples)

Design of a visual perception model with edge-adaptive Gabor filter and support vector machine for traffic sign detection

Traffic sign detection is a useful application for driving assistance systems, and it is necessary to accurately detect traffic signs before they can be identified. Sometimes, however, it is difficult to detect traffic sign, which may be obscured by other objects or affected by illumination or lightning reflections. Most previous work on this topic has been based on region of interest analysis using the color information of traffic signs. Although this provides a simple way to segment signs, this approach is weak when a sign is affected by illumination or its own color information is distorted. To overcome this, this paper introduces a robust traffic detection framework for cluttered scenes or complex city views that does not use color information. Moreover, the proposed method can detect traffic sign in the night. We establish an edge-adaptive Gabor function, which is derived from human visual perception. It is an enhanced version of the original Gabor filter, and filters out unnecessary information to provide robust recognition. It decomposes the directional information of objects and reflects specific shapes of traffic signs. Once the extracted feature is obtained, a support vector machine detects the traffic sign. Applying scale-space theory, it is possible to resolve the scaling problem of the objects that we want to find. Our system shows robust performance in traffic sign detection, and experiments on real-world scenes confirmed its properties.

A personalized Web search engine using fuzzy concept network with link structure

There has been much research on link-based search engines such as google and clever. They use link structure to find precision result. Usually, a link-based search engine produces higher-quality results than a text-based search engine. However, they have difficulty in producing the result fit to a specific users preference. Personalization is required to support a more appropriate result. Among many techniques, the fuzzy concept network based on a user profile can represent a users subjective interest properly. The paper presents another search engine that uses the fuzzy concept network to personalize the results from a link-based search method. The fuzzy concept network based on a user profile reorders five results of the link-based search engine, and the system provides a personalized high-quality result. Experimental results with three subjects indicate that the system developed searches not only relevant but also personalized Web pages on a users preference.

Automated synthesis of resilient and tamper-evident analog circuits without a single point of failure

This study focuses on the use of genetic programming to automate the design of robust analog circuits. We define two complementary types of failure modes: partial short-circuit and partial disconnect, and demonstrated novel circuits that are resilient across a spectrum of fault levels. In particular, we focus on designs that are uniformly robust, and unlike designs based on redundancy, do not have any single point of failure. We also explore the complementary problem of designing tamper-proof circuits that are highly sensitive to any change or variation in their operating conditions. We find that the number of components remains similar both for robust and standard circuits, suggesting that the robustness does not necessarily come at significant increased circuit complexity. A number of fitness criteria, including surrogate models and co-evolution were used to accelerate the evolutionary process. A variety of circuit types were tested, and the practicality of the generated solutions was verified by physically constructing the circuits and testing their physical robustness.

Evolutionary ensemble of diverse artificial neural networks using speciation

Recently, many researchers have designed neural network architectures with evolutionary algorithms but most of them have used only the fittest solution of the last generation. To better exploit information, an ensemble of individuals is a more promising choice because information that is derived from combining a set of classifiers might produce higher accuracy than merely using the information from the best classifier among them. One of the major factors for optimum accuracy is the diversity of the classifier set. In this paper, we present a method of generating diverse evolutionary neural networks through fitness sharing and then combining these networks by the behavior knowledge space method. Fitness sharing that shares resources if the distance between the individuals is smaller than the sharing radius is a representative speciation method, which produces diverse results than standard evolutionary algorithms that converge to only one solution. Especially, the proposed method calculates the distance between the individuals using average output, Pearson correlation and modified Kullback-Leibler entropy to enhance fitness sharing performance. In experiments with Australian credit card assessment, breast cancer, and diabetes in the UCI database, the proposed method performed better than not only the non-speciation method but also better than previously published methods.