• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.06a
• /
• pp.75-78
• /
• 2001

본 논문은 통계청에서 실시하는 인구주택 총조사로부터 획득된 각 개인의 직업 및 직종을 기술하고 있는 자연어를 입력받아 입력된 자연어가 의미하는 한국 표준 산업/구업 분류 코드의 후보들을 생성하는 산업/직업 코드 분류 도구를 제안한다. 코드 분류는 분류할 코드를 문서 범주로 간주하면 문서 분류와 동일한 문제로 생각할 수 있다. 하지만 본 산업/직업 코드 분류 문제는 입력되는 자연어의 길이가 한 두 문장 정도로 매우 짧아 문서 분류에 사용될 자질들이 개수가 주어 기존의 문서 분류 기법을 적용하기 어렵다. 이에 본 논문은 표준 코드를 기술하고 있는 내용을 미리 색인하고 입력된 자연어로부터 질의어를 생성하여 벡터공간모델로 질의어를 검색후 질의어와 일치율이 가장 높은 코드들을 분류될 후보 코드로 계시하는 정보검색 기법을 이용한 산업/직업 코드 분류 도구를 개발하였다.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.394-398
• /
• 2020

산업/직업분류 자동코딩시스템은 고용조사 등을 함에 있어 사업체 정보, 업무, 직급, 부서명 등 사용자의 다양한 입력을 표준 산업/직업분류에 맞춰 코드 정보를 제공해주는 시스템이다. 입력 데이터로부터 비지도학습 기반의 색인어 추출 모델을 학습하고, 부분단어 임베딩이 적용된 색인어 임베딩 모델을 통해 입력 벡터를 추출 후, 출력 분류 코드를 인코딩하여 지도학습 모델에서 학습하는 방법을 적용하였다. 기존 시스템의 분류 결과 데이터를 통해 대, 중, 소, 세분류에서 높은 정확도의 모델을 구축할 수 있으며, 기계학습 기술의 적용이 가능한 시스템임을 알 수 있다.

• 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 the Korean Society of Food Science and Nutrition
• /
• v.46 no.7
• /
• pp.876-885
• /
• 2017

This study aimed to identify the characteristics of Home Meal Replacement (HMR) product purchases and the need for HMR product development for Hong Kong consumers in order to suggest market segmentation strategies according to consumers' dietary lifestyle. For this, an online survey was conducted on a panel of 521 Hong Kong consumers with HMR purchase experience registered at a specialized organization. Data analysis was performed using SPSS (ver. 23.0). HMR purchase characteristics of Hong Kong consumers according to dietary lifestyle showed significant differences in all items, including 'number of purchases', 'purchase location', 'cost of single purchase', and 'reason for purchase'. According to dietary lifestyle, participants were divided into three clusters: 'High interest', 'normal interest', and 'low interest'. In the case of 'high interest in dietary life group', 'low-sodium food' was the most common, followed by 'heating food', 'low sugar food', and 'low calorie food'. In the case of 'moderate interest in dietary life group', 'low-sodium food' was the most common, followed by 'low sugar food', 'low calorie food', and 'nutritious meal'. In the case of 'low interest in dietary life group', 'low sugar food' was the most common, followed by 'low-sodium food', 'various new menu', and 'easy-to-carry dehydrated food'. For the 'high interest' group, the highest proportion of consumers were male in between the ages of 20 to 29, married, and worked in an office job. The 'high interest' consumers also showed a tendency to pay '15,000 to 20,000 KRW' per single purchase. The 'normal interest' group consisted of an even proportion of male and female consumers, with the most common age range being from 30 to 39 years, and most were married. These consumers preferred to spend 'less than 10,000 KRW' or '10,000 KRW to 15,000 KRW' per single purchase, which is in the lower price range for HMR purchases. The 'low interest in dietary life group' had more females gender-wise, were unmarried, and worked in an office job, For a single purchase, the 'low interest' group chose to pay less than 10,000 KRW, which is relatively lower than the other two clusters. The results of this study can be used as baseline data for building marketing strategies for HMR product development. It can also provide basic data and directions for new HMR export products that reflect consumer needs in order to create a market segmentation strategy for industrial applications.