Online shopping has revolutionized the way consumers make purchasing decisions. Customer reviews play a crucial role in this process, providing valuable insights into product quality and satisfaction. However, sifting through numerous reviews can be time-consuming for customers. This research aims to address this issue by proposing a novel method to generate a single summary for the overall reviews of each product across different years. The proposed approach utilizes the BART large xsum model to streamline the summarization process. Initially, the dataset is preprocessed to remove irrelevant reviews, such as those lacking product IDs or years. Subsequently, the dataset is divided into batches, each containing 1024 tokens, to facilitate efficient processing. Summaries are generated for each batch, and the output is then used as input for the next stage. Finally, a concise summary is produced for each product across different years. Experimental evaluations are conducted using the Amazon baby product reviews dataset, with the performance of the BART large xsum model assessed using confidence scores. The results demonstrate that the proposed method effectively generates concise summaries for the overall reviews of each product across different years, saving customers valuable time in decision-making.

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

Enhancing Product Review Understanding: Text Summarization with BART Large XSum

  • S. Gopika,
  • Mayuri Mahimaa Balaji,
  • M. Vishwanath,
  • K. Karthikayani

摘要

Online shopping has revolutionized the way consumers make purchasing decisions. Customer reviews play a crucial role in this process, providing valuable insights into product quality and satisfaction. However, sifting through numerous reviews can be time-consuming for customers. This research aims to address this issue by proposing a novel method to generate a single summary for the overall reviews of each product across different years. The proposed approach utilizes the BART large xsum model to streamline the summarization process. Initially, the dataset is preprocessed to remove irrelevant reviews, such as those lacking product IDs or years. Subsequently, the dataset is divided into batches, each containing 1024 tokens, to facilitate efficient processing. Summaries are generated for each batch, and the output is then used as input for the next stage. Finally, a concise summary is produced for each product across different years. Experimental evaluations are conducted using the Amazon baby product reviews dataset, with the performance of the BART large xsum model assessed using confidence scores. The results demonstrate that the proposed method effectively generates concise summaries for the overall reviews of each product across different years, saving customers valuable time in decision-making.