Stephen Kent A. Malagday, Mark Raphael V. Sto. Domingo.
Named entity recognition on E-Mono: An algorithm enhancement applied in sentiment analysis - Undergraduate Thesis: (Bachelor of Science in Computer Science) Pamantasan ng Lungsod ng Maynila. 2023.
ABSTRACT: Sentiment analysis is a critical component of natural language processing that seeks to classify emotions conveyed in the text. To achieve this goal, various approaches have been developed, and one commonly used method is the Extended Max- Occurrence with Normalized Non-Occurrence (EMONO) term weighting scheme. The EMONO scheme builds upon the Max-Occurrence with a Normalized Non-Occurrence (MONO) approach, which considers the frequencies of term occurrences in sentiment classes. However, the original EMONO approach has a limitation in that it does not consider the importance of named entities and their associated sentiment. Recognizing this gap, the researchers proposed an enhanced approach that integrates the E-MONO term weighting scheme with Name Entity Recognition (NER). By incorporating NER, the researchers aim to enhance the accuracy of sentiment analysis by identifying named entities and analysing the sentiment expressed towards specific entities. The combination of NER and the enhanced EMONO term weighting scheme aims to capture the nuanced sentiment expressed towards named entities, resulting in improved sentiment analysis outcomes. In experimental results, the proposed approach achieved an accuracy rate of 84% in classifying sentiment using E-MONO with the integrated Named Entity Recognition. These findings demonstrate the effectiveness of the combined approach in accurately identifying and analysing sentiment towards named entities, contributing to more precise sentiment analysis results overall.
5
5
academic writing
QA76.9.A43 M34 2023
Named entity recognition on E-Mono: An algorithm enhancement applied in sentiment analysis - Undergraduate Thesis: (Bachelor of Science in Computer Science) Pamantasan ng Lungsod ng Maynila. 2023.
ABSTRACT: Sentiment analysis is a critical component of natural language processing that seeks to classify emotions conveyed in the text. To achieve this goal, various approaches have been developed, and one commonly used method is the Extended Max- Occurrence with Normalized Non-Occurrence (EMONO) term weighting scheme. The EMONO scheme builds upon the Max-Occurrence with a Normalized Non-Occurrence (MONO) approach, which considers the frequencies of term occurrences in sentiment classes. However, the original EMONO approach has a limitation in that it does not consider the importance of named entities and their associated sentiment. Recognizing this gap, the researchers proposed an enhanced approach that integrates the E-MONO term weighting scheme with Name Entity Recognition (NER). By incorporating NER, the researchers aim to enhance the accuracy of sentiment analysis by identifying named entities and analysing the sentiment expressed towards specific entities. The combination of NER and the enhanced EMONO term weighting scheme aims to capture the nuanced sentiment expressed towards named entities, resulting in improved sentiment analysis outcomes. In experimental results, the proposed approach achieved an accuracy rate of 84% in classifying sentiment using E-MONO with the integrated Named Entity Recognition. These findings demonstrate the effectiveness of the combined approach in accurately identifying and analysing sentiment towards named entities, contributing to more precise sentiment analysis results overall.
5
5
academic writing
QA76.9.A43 M34 2023