Text Sentiment Analysis for Movie Theme Forecasting Based on Natural Language Processing

Authors

  • Xuhui Song

DOI:

https://doi.org/10.54097/2w33re40

Keywords:

Python, Web Crawling, Movie Themes, Douban, NLP.

Abstract

With the advancement of the internet, online film reviews have emerged as a significant platform for audiences to share their viewing experiences. This study, leveraging short review data from Douban movies, proposes a film genre prediction method that integrates text sentiment analysis and high-frequency word matching. By scraping short reviews from Douban's Top 250 films and utilizing Python's jieba package for word segmentation, high-frequency terms were extracted and matched with genres using a constructed sentiment dictionary. Experimental results demonstrate that this approach effectively identifies mainstream genres such as drama and romance, though its performance in detecting niche genres requires further enhancement. The key innovation lies in directly extracting genre-related features from audience reviews and improving keyword recognition accuracy through sentiment analysis. While the method proves effective to a certain extent in predicting film genres, there remains room for improvement in both accuracy and generalizability. This study not only introduces a novel technical framework for film genre analysis but also provides empirical insights into the relationship between audience genre preferences and emotional responses. Future research could explore the integration of deep learning models to further enhance the precision and robustness of genre prediction.

Downloads

Download data is not yet available.

References

[1] Yu Yang. Sentiment analysis and topic extraction research on Douban movie reviews. Yunnan University of Finance and Economics, 2018.

[2] Ma Jia, Liu Yubang. Text mining and visualization data analysis of Douban short reviews on the program “Planting Season”: Research based on LAC sentiment analysis model. New Legend, 2025, (20): 73-75.

[3] Zhao Yanyan, Qin Bing, Liu Ting. Text sentiment analysis. Journal of Software, 2010, 21 (08): 1834-1848.

[4] Lu Wenxing, Wang Yanfei. A review of Chinese text sentiment analysis research. Application Research of Computers, 2012, 29 (06): 2014-2017.

[5] Liu Yulin, Jian Lirong. E-commerce online review data mining based on text sentiment analysis. Statistics and Information Forum, 2018, 33 (12): 119-124.

[6] Zhang Lu. Sentiment analysis of overseas readers’ reception and evaluation of Chinese translated literature: A case study of the English version of The Three-Body Problem based on Python sentiment analysis. Foreign Language Research, 2019, 36 (04): 80-86.

[7] Yan Chengxiao. Short text opinion extraction and polarity word analysis based on NLP. Software, 2022, 43 (03): 121-123.

[8] Qu Xiaoxiao. Research on clustering analysis and visualization methods of movie review data. Shandong University, 2018.

[9] Xu Delong, Lin Min, Wang Yurong, et al. A review of NLP data augmentation methods based on large language models. Journal of Frontiers of Computer Science and Technology, 2025, 19 (06): 1395-1413.

[10] Liu Zhiming, Liu Lu. Empirical study on Chinese microblog sentiment classification based on machine learning. Computer Engineering and Applications, 2012, 48 (01): 1-4.

[11] Zhou Zhonghua, Zhang Huiran, Xie Jiang. Sina Weibo data crawler based on Python. Journal of Computer Applications, 2014, 34 (11): 3131-3134.

Downloads

Published

29-01-2026

Issue

Section

Articles