Comparing ELM with SVM in the field of sentiment classification of social media text data

• Main Authors: CHEN, Zhihuan, WANG, Zhaoxia, LIN, Zhiping, YANG, Ting
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2018
• Subjects: ELM; SVM; Sentiment Classification; Social Media; Learningbased Method; Artificial Intelligence and Robotics; Databases and Information Systems
• Online Access: https://ink.library.smu.edu.sg/sis_research/6121; https://ink.library.smu.edu.sg/context/sis_research/article/7124/viewcontent/2020_Compare_SVM_ELM_13th_version_27_11_2018_ELM_2017.pdf
• Institution: Singapore Management University

Machine learning has been used in various fields with thousands of applications. Extreme learning machine (ELM), which is the most recently developed machine learning algorithm, has become increasingly popular for its good generalization ability. However, it has been relatively less applied to the domain of social media. Support Vector Machine (SVM), another popular learning-based algorithm, has been applied for sentiment classification of social media text data and has obtained good results. This paper investigates and compares the capabilities of these two learning-based methods in the field of sentiment classification of social media. The results indicate that SVM can obtain good performance for analyzing small datasets, while for large datasets, ELM performs better than SVM. This research also indicates that ELM has the potential application in the domain of social media analysis.