Anticipate Movie Theme from Subtitle: A Deep Learning Approach

In the age of big data, prediction is crucial to making accurate assumptions about connected data. People must use other relevant data to determine the proper data relationships from other data sources. One area where movies may be predicted and categorized using subtitles is the movie genre. Variou...

Full description

Saved in:
Bibliographic Details
Published in2024 47th International Conference on Telecommunications and Signal Processing (TSP) pp. 205 - 210
Main Authors Alzoubi, Yehia Ibrahim, Topcu, Ahmet E., Elbasi, Ersin, Buyukyilmaz, Mucahit, Cibikdiken, Ali Osman
Format Conference Proceeding
LanguageEnglish
Published IEEE 10.07.2024
Subjects
Accuracy
Deep learning
genre
movie
Prediction methods
Signal processing
Soft sensors
Text categorization
text classification
Online AccessGet full text

Cover

More Information
Summary:In the age of big data, prediction is crucial to making accurate assumptions about connected data. People must use other relevant data to determine the proper data relationships from other data sources. One area where movies may be predicted and categorized using subtitles is the movie genre. Various data models are used in other movie prediction methods to address the issue. This study uses Deep Learning (DL) techniques to infer the movie genre from subtitle data. The OpenSubtitle and IMDb websites provided the dataset for the study. This collection includes XML-formatted subtitle files for various films, series, and related film genres. Each subtitle was turned into a 100,000-dimensional vector as part of the preprocessing of the subtitle files. The 5-fold cross-validation methodology was used, in this study, to evaluate the DL-Long Short-Term Memory (LSTM) model, and the results were quantified and presented using various techniques. The accuracy rating for the proposed model was 93.97%.
ISSN:2768-3311
DOI:10.1109/TSP63128.2024.10605925