지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학술대회자료
- 저자정보
- 발행연도
- 2024.10
- 수록면
- 67 - 70 (4page)
이용수
초록· 키워드
오류제보하기
Named Entity Recognition (NER) is a theoretical concept and a practical tool with significant applications in the food industry. Our research, which is novel in its approach, focuses on extracting and categorizing food items, ingredients, nutritional information, allergens, preparation methods, quantities, and more from unstructured text data like recipes and product descriptions. This capability supports personalized nutrition, food safety, and regulatory compliance by automatically identifying and classifying critical information. We developed a Bidirectional Long Short-Term Memory (BiLSTM) model using pre-trained GloVe (Global Vectors for Word Representation) embeddings to capture contextual relationships in food-related texts. We then introduced a custom-weighted categorical cross-entropy loss function to improve the recognition of less frequent entities. Enhancing this model further, we added an attention mechanism, which dynamically focuses on essential words and captures complex relationships, thereby improving the model’s performance. To optimize the model, we introduced hyperparameter tuning and K-fold cross-validation. Our final architecture, a unique combination of a DenseNet-inspired dense block with BiLSTM and attention layers with convolutional operations for advanced feature extraction, achieved a high F1- score of 99.66%, bypassing state-of-the-art by a significant margin. This progression from a basic BiLSTM to an advanced model improves the accuracy and reliability of NER in the food industry, supporting various practical applications like food safety, policy compliance, dietary assessment, and personalized nutrition.
목차
ABSTRACT
Ⅰ. Introduction
Ⅱ. Methodology
Ⅲ. Experimental Results
Ⅴ. Conclusion
References
참고문헌 (0)
참고문헌 신청함께 읽어보면 좋을 논문
논문 유사도에 따라 DBpia 가 추천하는 논문입니다. 함께 보면 좋을 연관 논문을 확인해보세요!
-
Encoding Dictionary Feature for Deep Learning-based Named Entity Recognition
International JOURNAL OF CONTENTS
2021 .12
-
A Study on the Performance Analysis of Entity Name Recognition Techniques Using Korean Patent Literature
JOURNAL OF ADVANCED INFORMATION TECHNOLOGY AND CONVERGENCE
2020 .12
-
BiLSTM을 사용한 개선된 타켓 트랙킹 연구
한국항공우주학회 학술발표회 초록집
2021 .07
-
Evaluating and applying deep learning-based multilingual named entity recognition
Journal of Advanced Marine Engineering and Technology (JAMET)
2018 .02
-
Cross-Domain Text Sentiment Classification Method Based on the CNN-BiLSTM-TE Model
JIPS(Journal of Information Processing Systems)
2021 .08
-
Transfer Learning 기반 BiLSTM을 통한 밀링공정 절삭력 예측
한국정밀공학회 학술발표대회 논문집
2020 .09
-
A Study on Sales Prediction Model Based on BiLSTM-GAT Using Credit Card Transaction Data
Journal of KIISE
2024 .09
-
BiLSTM모델을 이용한 사용자 평균 평점을 통한 온라인 후기 평점 예측
Proceedings of KIIT Conference
2024 .05
-
데이터 증강 기법을 이용한 한글 개체명 인식
대한산업공학회지
2022 .04
-
데이터 증강 기법을 이용한 한글 개체명 인식
대한산업공학회 추계학술대회 논문집
2021 .11
-
An Attention-Based CNN-BiLSTM Model for Korean Voice Phishing Detection
한국정보과학회 학술발표논문집
2022 .06
-
CNN과 BiLSTM을 이용한 한국어 텍스트 분류 모델의 성능 비교: 보이스 피싱 대화 데이터 기반
Proceedings of KIIT Conference
2023 .06
-
Optical Korean Character Recognition using CRNN with BiLSTM for video analysis
대한산업공학회 춘계공동학술대회 논문집
2022 .06
-
모바일 엣지 컴퓨팅(MEC)을 위한 BiLSTM을 이용한 심층 강화 학습 기반 태스크 오프로딩 알고리즘
멀티미디어학회논문지
2024 .08
-
Flexibility Availability Forecasting of Distributed Energy Resources for Virtual Power Plant using Hybrid CNN-BiLSTM Network
대한전기학회 학술대회 논문집
2024 .07
-
BiLSTM을 이용한 SDI(하천유량가뭄지수)예측
대한토목학회 학술대회
2024 .10
-
Comparative study of text representation and learning for Persian named entity recognition
[ETRI] ETRI Journal
2022 .10
-
어텐션 기반의 BiLSTM-AE 모델을 이용한 철도차량 공기압축기 이상탐지
제어로봇시스템학회 국내학술대회 논문집
2023 .06
-
DFE 기반 복층 BiLSTM Equalizer 성능 검증
한국통신학회 학술대회논문집
2022 .06
-
Ship Motion-Based Prediction of Damage Locations Using Bidirectional Long Short-Term Memory
한국해양공학회지
2022 .10
최근 본 자료
댓글(0)
0