Framework for emoji, sentiment and emotion-aided cyber bullying detection in Hindi-English, which provides benchmark Hindi-English code-mixed corpus annotated with bully, sentiment and emotion labels to determine whether sentiment and emotion label information assists in identifying cyber bully
	MOHITH K; KUMAR M P; PAVANI D; NAVEEN V; MOUNIKA C
	2025-03-29
专利权人	MOHITH K (MOHI-Individual) ; KUMAR M P (KUMA-Individual) ; PAVANI D (PAVA-Individual) ; NAVEEN V (NAVE-Individual) ; MOUNIKA C (MOUN-Individual)
申请日期	2025-03-29
专利号	IN202541030914-A
成果简介	NOVELTY - The framework provides a benchmark Hindi-English code-mixed corpus annotated with bully, sentiment and emotion labels to determine whether sentiment and emotion label information assists in identifying cyber bully with more accuracy. The main task-cyberbully detection (CBD), an attention-based multitask multimodal framework are used for detectng sentiment and emotion-aided cyber bullying. The IA and IRA are incorporated into the model for the efficient representations of different modalities and helping to learn generalized features over multiple tasks. The main task-CBD framework outperforms all single-task and unimodal, with a significant margin. An attention-based multimodal, adversarial multitasking framework is proposed for CBD considering two auxiliary tasks such as sentiment analysis (SA) and emotion recognition (ER). USE - Framework for emoji, sentiment and emotion-aided cyber bullying detection in Hindi-English in social networking platforms. ADVANTAGE - The experimental results of the framework demonstrate that concurrent execution of three tasks such as CBD, SA, and ER improves the performance of the main task-CBD. The multimodal framework outperforms the unimodal framework by a substantial margin since emoji data add some extra information to inputs, which helps the classifier for better prediction. All the reported results for the proposed model and baselines are statistically significant. DESCRIPTION OF DRAWING(S) - The drawing shows a flow diagram of the framework for emoji, sentiment and emotion-aided cyber bullying detection in Hindi-English.
IPC 分类号	G06F-040/00 ; G06F-040/30 ; G06Q-050/00
国家	印度
专业领域	信息技术
语种	英语
成果类型	专利
文献类型	科技成果
条目标识符	http://119.78.100.226:8889/handle/3KE4DYBR/13359
专题	中国科学院新疆生态与地理研究所
作者单位	1.MOHITH K (MOHI-Individual) 2.KUMAR M P (KUMA-Individual) 3.PAVANI D (PAVA-Individual) 4.NAVEEN V (NAVE-Individual) 5.MOUNIKA C (MOUN-Individual)
推荐引用方式 GB/T 7714	MOHITH K,KUMAR M P,PAVANI D,et al. Framework for emoji, sentiment and emotion-aided cyber bullying detection in Hindi-English, which provides benchmark Hindi-English code-mixed corpus annotated with bully, sentiment and emotion labels to determine whether sentiment and emotion label information assists in identifying cyber bully. IN202541030914-A[P]. 2025.

条目包含的文件
条目无相关文件。

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。