Multimodal Sentiment Analysis for Urdu Language

0 ratings - Please login to submit your rating.


The "Multi-modal Sentiment Analysis Dataset for Urdu Language Opinion Videos" is a valuable resource aimed at advancing research in sentiment analysis, natural language processing, and multimedia content understanding. This dataset is specifically curated to cater to the unique context of Urdu language opinion videos, a dynamic and influential content category in the digital landscape.

Dataset Description:

  • Size and Diversity: This dataset comprises an extensive collection of Urdu language opinion videos, encompassing a wide spectrum of topics and sentiments. It consists of a total of 214 videos, each of varying lengths, offering a diverse and comprehensive representation of the Urdu language content landscape.
  • Sentiment Annotations: The dataset is meticulously annotated with sentiment labels, providing information on the emotional tone expressed in each video. The sentiment labels include "positive," "negative," and "neutral," offering a nuanced understanding of the sentiment conveyed in these multimedia opinion pieces.
  • Multi-modal Approach: A unique feature of this dataset is its multi-modal approach. It combines text, audio, and visual data to enable researchers to delve into the various dimensions of sentiment analysis within the context of opinion videos. The multi-modal annotations encompass the textual content of spoken words, the auditory characteristics of the videos, and the visual cues from the video frames.

Significance and Applications:

This dataset holds significant value for both the research community and practical applications:

  • Research Advancement: Researchers can employ this dataset to investigate the complex landscape of sentiment analysis within the context of opinion videos. It facilitates inquiries into sentiment trends, the development of sentiment analysis models, and the creation of sentiment-aware multimedia content analysis tools.
  • Content Recommendation: The dataset can play a pivotal role in the development of content recommendation systems that cater to the emotional preferences of viewers. Understanding sentiment in opinion videos is crucial for improving content engagement and user experience.
  • User Engagement Analysis: The dataset can empower studies focused on user engagement and interaction with multimedia content. It is an essential resource for researchers aiming to decode the factors that influence viewer reactions and engagement in a multimedia context.

In summary, the "Multi-modal Sentiment Analysis Dataset for Urdu Language Opinion Videos" offers a unique, comprehensive, and diverse resource for researchers interested in exploring the dynamic world of sentiment within the context of Urdu opinion videos. Its multi-modal annotations and large dataset size provide a robust foundation for advancing research in this domain and developing innovative multimedia analysis solutions.

Researchers are encouraged to explore and utilize this dataset for various academic and commercial purposes, fostering innovation in sentiment analysis and multimedia understanding. The dataset is made available with open access to facilitate collaborative research and to contribute to the broader knowledge in the field.




The Dataset consists of the following items:

  • Full-length videos
  • segmented videos
  • segmented audios
  • text transcription


Due to limited space full dataset is not uploaded at dataport. For interested users the dataset link is provided.

Submitted by Ghulam Rabbani on Wed, 10/18/2023 - 02:04