Scholarlink

<p>Bengali, one of the most widely spoken yet historically marginalized languages, presents substantial challenges for machine translation due to its extensive vocabulary. Recent advances in neural machine translation, particularly the introduction of T5-Transformers, have significantly improved Bengali-to-English translation quality Model performance is evaluated using two standard metrics, BLEU and ROUGE-L, which assess the similarity between predicted and reference translations. An experimental result demonstrates strong performance, achieving BLEU and ROUGE-L scores of 75 and 78, respectively. This study investigates the effectiveness of combining four publicly available datasets on cricket to enhance sentiment analysis accuracy by integrating Long Short-term Memory (LSTM) networks- a recurrent neural architecture well suited for sequence modeling-within an encoder-decoder framework. These findings highlight the potential of advanced neural techniques, including T5-Transformers and LSTM, to deliver highly accurate and efficient Bengali-to-English translations with effective sentiment analysis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Sentiment Analysis on Converting Marginalized Language to English Language

  • 期刊论文

  • 发表日期：

    2026年6月23日

查看全文
  • V. Vani,
  • B. Pabitha,
  • Chandra Praba,
  • Mahanth Kumar Addanki,
  • Srujana Jawalkar,
  • M. Priyangka

摘要

Bengali, one of the most widely spoken yet historically marginalized languages, presents substantial challenges for machine translation due to its extensive vocabulary. Recent advances in neural machine translation, particularly the introduction of T5-Transformers, have significantly improved Bengali-to-English translation quality Model performance is evaluated using two standard metrics, BLEU and ROUGE-L, which assess the similarity between predicted and reference translations. An experimental result demonstrates strong performance, achieving BLEU and ROUGE-L scores of 75 and 78, respectively. This study investigates the effectiveness of combining four publicly available datasets on cricket to enhance sentiment analysis accuracy by integrating Long Short-term Memory (LSTM) networks- a recurrent neural architecture well suited for sequence modeling-within an encoder-decoder framework. These findings highlight the potential of advanced neural techniques, including T5-Transformers and LSTM, to deliver highly accurate and efficient Bengali-to-English translations with effective sentiment analysis.