• Journal of the Korea Society of Computer and Information
• /
• v.28 no.4
• /
• pp.21-29
• /
• 2023

In this paper, we propose an AI-based system for automatically classifying industry and occupation codes in the population census. The accurate classification of industry and occupation codes is crucial for informing policy decisions, allocating resources, and conducting research. However, this task has traditionally been performed by human coders, which is time-consuming, resource-intensive, and prone to errors. Our system represents a significant improvement over the existing rule-based system used by the statistics agency, which relies on user-entered data for code classification. In this paper, we trained and evaluated several models, and developed an ensemble model that achieved an 86.76% match accuracy in industry and 81.84% in occupation, outperforming the best individual model. Additionally, we propose process improvement work based on the classification probability results of the model. Our proposed method utilizes an ensemble model that combines transfer learning techniques with pre-trained models. In this paper, we demonstrate the potential for AI-based systems to improve the accuracy and efficiency of population census data classification. By automating this process with AI, we can achieve more accurate and consistent results while reducing the workload on agency staff.

• Journal of KIISE:Software and Applications
• /
• v.27 no.6
• /
• pp.607-617
• /
• 2000

Concurrent programs, which are composed of threads that can operate concurrently, are intrinsically nondeterministic. Therefore, we can not directly apply classical testing or debugging techniques developed for sequential programs to concurrent programs because of the nonreproducibility due to nondeterminism. In this paper, we present a source tranformation method to guarantee the reproducibility of multi-threaded programs written in Java programming language. Once a mutli-threaded program has been transformed according to the transformation rules, we can force the program's execution to follow the given sequence of synchronized methods repeatedly and reduce the efforts to find the source of the errors. In addition, we present a method for checking the feasibility of the synchronized method sequence to be replayed.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.11 no.2
• /
• pp.1-10
• /
• 2015

As programs get more complicated and they are developed by various hands, the possibility that there are program bugs in the code has been increasing. And developers usually run unit tests to find these problems in the code. Besides, the developers are at the pain of getting stability of the code when they have to modify a code very often for clients requirements. In the methodlogy of TDD(Test Driven Development), developers write a unit test code first, and then write a program code for passing the unit test. The unit test must include the boundary condition test the reason why the possibility of occurring the bugs is very high. When failed to pass the test because of the value of a function is incorrect, not existed, out of the range or not matched etc, the program code will return the error code or occur the exception. In the document, the system is designed and implemented in order to insert the generated code automatically or suggest it to the developer, when the boundary condition test is failed. In conclusion, it is possible that the developer will get the code stability by searching the code and checking the code to be omitted automatically through this system.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.8 no.6
• /
• pp.63-69
• /
• 2008

In this paper, our goal is to find a power-effective protocol that improves the accuracy of transmission in sensor networks. Therefore we propose a cooperative communication protocol based on MRC(Maximal Ratio Combining) and cyclic code. In our proposal, one sensor node assists two others to communicate with a clusterhead that can get diversity effect and MRC can improve diversity effect also. The proposed protocol with cyclic code can correct error up to 3-bit and reduce decoding complexity compared with convolutional code. Simulation results reveal proposed protocol can save the network energy up to 6dB over single-hop protocol at BER(Bit Error Rate) of $10^{-2}$.

• Journal of Cadastre & Land InformatiX
• /
• v.44 no.1
• /
• pp.165-177
• /
• 2014

The KML verification tool supports people who want to produce the highest quality KML file. In other words, it validate that a given KML document is well-formed with respect to XML standard meaning, and conform not only to the KML schema and the specification. Then it's only to notify error code line. People who want to use the KML file written by others would like to know both whether the validity of that file and general summary of feature's location, shape, and number. In this study, we recommended the user-friendly KML validator using OpenLayers and reporting geometries and images of the KML file.

• The Journal of Korean Association of Computer Education
• /
• v.9 no.4
• /
• pp.55-61
• /
• 2006

The Personal Software Process(PSP) is a rigorous, measurements based process designed to help software developers produce high-quality codes. Learning PSP requires substantial and tedious record keeping including clerical computations with plenty of data occurred at every project. The required task reduces the educational effect and motive by time consuming characteristic and easy occurrence of errors. To alleviate these problems, a web-based PSP log system having streamlined mechanism of data manipulation is designed and implemented at this study. Three years of experiment shows that the system not only improves productivity and accuracy of the process but also allows students and instructors to focus on learning and teaching PSP, rather than PSP data collection.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06b
• /
• pp.401-404
• /
• 2011

운영체제와 같은 대규모 프로그램은 확장을 용이하게 하기 위해 장치 드라이버를 모듈 구조를 통해 동작시킨다. 그러나 이러한 드라이버 모듈은 커널과 분리되어 개발되기 때문에 상대적인 취약성을 가지고 있다. 이러한 취약성으로 인해 발생한 고장은 드라이버뿐만 아니라 커널 전체에 오류를 야기해 시스템의 안정성을 저하시키는 요인이 된다. 이러한 문제를 해결하기 위해 안전하지 않은 드라이버를 커널로부터 격리하고자 하는 논의가 진행되었고, 이후 많은 연구를 통해 여러 가지 접근 방식을 통한 해결방법이 제안되었다. 그러나 기존의 방법은 드라이버 소스코드의 수정이 필요하거나 드라이버 요청에 대한 응답속도가 매우 느린 현실적인 제약이 있었다. 본 논문에서는 기존방법의 한계를 극복하는 동적 컴파일러를 이용한 고장 분리 기법을 제안한다. 본 제안 기법은 기존 드라이버를 그대로 사용가능하며 기존 기법의 분리 정책을 모두 적용할 수 있다. 리눅스 커널에 본 제안기법을 적용하여 동적 컴파일러의 성능을 측정하고 본 제안 기법의 특성을 기존 기법과 비교함으로써 본 제안기법이 매우 적은 부하만으로 장치 드라이버의 고장 분리가 가능함을 확인하였다.

• The KIPS Transactions:PartA
• /
• v.12A no.1 s.91
• /
• pp.1-6
• /
• 2005

This paper presents a new algorithm on low power survivor path memory implementation of the trace-back systolic array Viterbi algorithm. A novel idea is to reuse the already-generated trace-back routes to reduce the number of trace-back operations. And the spurious switching activity of the trace-back unit is reduced by making use of a clock gating method. Using the SYNOPSYS power estimation tool, DesignPower, our experimental result shows the average $40{\%}$ power reduction and $23{\%}$ area increase against the trace-back unit introduced in [1].

• Journal of Internet Computing and Services
• /
• v.6 no.1
• /
• pp.51-64
• /
• 2005

In this paper, we introduce a system for evaluating similarity of C program source code using method which compares syntax-trees each others. This method supposes two characteristic features as against other systems. It is not sensitive for program style such as indentation, white space, and comments, and changing order of control structure like sentences, code block, procedures, and so on. Another is that it can detect a syntax-error cause of using paring technique, We introduce algorithms for similarity evaluation method and grouping method that reduces the number of comparison, In the examination section, we show a test result of program similarity evaluation and its reduced iteration by grouping algorithm.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.4
• /
• pp.735-743
• /
• 2020

Secure coding is a technique that detects malicious attack or unexpected errors to make software systems resilient against such circumstances. In many cases secure coding relies on static analysis tools to find vulnerable patterns and contaminated data in advance. However, secure coding has the disadvantage of being dependent on rule-sets, and accurate diagnosis is difficult as the complexity of static analysis tools increases. In order to support secure coding, we apply machine learning techniques, such as DNN, CNN and RNN to investigate into finding major weakness patterns shown in secure development coding guides and present machine learning models and experimental results. We believe that machine learning techniques can support detecting security weakness along with static analysis techniques.