Year: 2023
Type: article
Abstract: This paper analyzes Instagram text sentiment to explore people's sentimental engagement in social platforms.First, data mining technology is used to obtain the Instagram text information, and the Instagram text dataset is obtained through processing.Second, the deep learning technology in Machine Learning (ML) and the word embedding technology of Natural Language Processing are used to construct the Instagram text classification model, including the construction of the model structure and the specific design of each network layer.Finally, the set experiment verifies the effect of the constructed text sentiment analysis model.The verification results show that: (1) The sentiment classification model constructed here can effectively extract the characteristics of long sentences while considering the article's word order features.The values of each indicator are also greater than 79%, and all are higher than other models.(2) The most concentrated range of Instagram text sentiment values in 2022 was [0.53, 0.59], indicating that Instagram's text sentiment was positive this year on average.This paper aims to provide a theoretical basis for further ML development in text sentiment analysis.
DOI: 10.1145/3606370
Authors: Chia-Pang Chan, Jun-He Yang
Institution: Cheng Shiu University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4382987285