Izvorni znanstveni članak

https://doi.org/10.17559/TV-20230629000774

Construction of an Inquiry Letter Sentiment Dictionary Using SO-PMI and Word2Vec for Sentiment Analysis

Wei Wang ; School of Economics & Management, University of Science and Technology Beijing, No. 30 Xueyuan Road Haidian District, Beijing 100083, China
Guiying Wei ; School of Economics & Management, University of Science and Technology Beijing, No. 30 Xueyuan Road Haidian District, Beijing 100083, China
Sen Wu ; School of Economics & Management, University of Science and Technology Beijing, No. 30 Xueyuan Road Haidian District, Beijing 100083, China *
Huixia He ; School of Economics & Management, University of Science and Technology Beijing, No. 30 Xueyuan Road Haidian District, Beijing 100083, China

* Dopisni autor.

Sažetak

Sentiment analysis of financial texts provides valuable insights into market dynamics, but relies on domain-specific dictionaries. Inquiry letters as particular financial texts lack tailored sentiment dictionaries, limiting research in the important domain. This study develops an Inquiry Letter Sentiment Dictionary (ILSD) using innovative integration of semantic orientation pointwise mutual information (SO-PMI), word2vec, and manual screening. The ILSD leverages co-occurrence and contextual information to expand coverage and sentiment word capture capabilities beyond existing general and financial dictionaries. We collected 1754 inquiry letters from the Shanghai Stock Exchange (SSE). The experiment results indicate that the ILSD performs better than other sentiment dictionaries according to coverage, accuracy (ACC), F1 weighted score (F1Weighted), Matthews correlation (MCC), and geometric mean (G-mean), which proves the effectiveness of the ILSD in practice.

Ključne riječi

inquiry letter; sentiment classification; sentiment dictionary; SO-PMI; word2vec

Hrčak ID:

325847

URI

https://hrcak.srce.hr/325847

Datum izdavanja:

31.12.2024.

Posjeta: 10 *