International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 12 | Dec 2024
p-ISSN: 2395-0072
www.irjet.net
ENHANCING HOPE SPEECH DETECTION IN SOCIAL MEDIA USING TRANSFORMER MODELS Aishwarya V. Nayak1 1PG Student, Dept. of CSE Engineering, VTU University, Karnataka, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The proliferation of online communication has
speech" might demotivate the speaker and reduce the importance of the topic. It important to recognize them in order to make an informed choice, given the greater effect of comments that express hope versus those that do not in social media settings.
led to an increase in both positive and negative content. While considerable efforts have been directed toward detecting and mitigating hate speech, the identification and promotion of hope speech, which has the potential to foster positive social change, has garnered less attention. This study explores the application of various machine learning and deep learning models to detect hope speech in online textual content. The models evaluated include Logistic Regression (LR), Support Vector Machine (SVM), Bidirectional Encoder Representations from Transformers (BERT), and Recurrent Neural Networks (RNN). The models' performance was measured using metrics such as accuracy, precision, recall, and F1-score. BERT and RNN models outperformed LR and SVM, emphasizing the importance of contextual understanding in hope speech detection. This study demonstrates that deep learning models, particularly BERT, outperform traditional machine learning methods in detecting hope speech.
Hope speech includes sayings that uplift, encourage, and assist people while promoting positivity and a sense of community. Detecting hope speech can benefit applications such as increasing mental health support, improving the overall ambiance of online platforms, and creating more engaging and supportive environments [2]. This investigates the concept of hope speech, the difficulties in identifying it, and techniques for anticipating hope speech.
1.1 Importance of Hope Speech
1.INTRODUCTION
A hope speech is a type of speech intended to uplift, encourage, and assist people or groups. Hope speech, in contrast to other positive communication techniques, is concentrated on fostering a feeling of hope and optimism.[3] These speeches work well in a range of contexts, such as social movements, crisis communication, and mental health assistance.
Social media sites and online forums have replaced traditional ways of communication for millions of people globally in the current digital era. Social media provides a virtual platform for people to develop, exchange, and work with each other on ideas, facts, and opinions [1]. Even though English is the most widely used language on social media platforms, people from various linguistic backgrounds connect and exchange ideas in their native tongues.
In order to lessen the negative effects of poisonous content on the internet, hope speech is crucial. People's mental health can be greatly impacted by supportive and community-building relationships. [2] Online platforms may become safer, friendlier places where people feel appreciated and supported by encouraging hope speech. Furthermore, hope speech has the power to favorably affect society attitudes and behaviors.
Social media provides deep insights into people's behaviour and is a valuable source of scientific research on topics pertaining to Natural Language Processing (NLP). We exchange our experiences, including those that are necessary for everyone to have during difficult times, such as gratitude for achievements, joy, anger, sadness, and inspiration from defeats [1]. These remarks, sometimes known as "hope speech," convey the well-being of the speaker. "Not hope speech" refers to the other sorts of remarks that demoralize, mistreat, or discourage someone based on their gender, race, handicap, or other minority. The positive recommendations from other viewers improve the video's coherence or substance.[1] On the other hand, remarks like "not hope
Regardless of its importance, detecting hope speech poses special obstacles. Unlike harmful material, which may be easily classified based on negativity or aggressiveness, hope speech is distinguished by its constructive and uplifting quality, making it more nuanced and context-dependent.[3] Detecting and distinguishing hope speech necessitates sophisticated procedures and a comprehension of context, tone, and intent.
Key Words: Hope Speech Detection, Machine Learning, Deep Learning, LR, SVM, BERT, RNN, TF-IDF, GloVe, Word2Vec and BERT Embeddings.
© 2024, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 655