Scholarlink

<p>The advent of the machine learning era means increasing data volumes and real-time data processing requirements especially in the medical field to improve the reliability of treatment results. Traditional data indexing has requirements on the storage space of medical devices and does not meet the characteristics of lightweight smart devices. In this paper, we propose a multi-dimensional medical data learned index, design novel mapping function for multi-dimensional data processing, and index multi-dimensional data using segmented linear function. It can efficiently perform data access and indexing processing on lightweight intelligent data processing devices and transmit abnormal data to medical centres. After a series of experiments, the results show that the proposed index supports direct data processing on lightweight intelligent data processing devices, which outperforms the traditional index by more than 10 times.</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

Intelligent Indexing of Medical Data Based on Machine Learning

  • 期刊论文

  • 发表日期：

    2025年6月18日

查看全文
  • Lijiao Sun,
  • Desheng Chen,
  • Shuai Xiao,
  • Meng Xi,
  • Jiabao Wen

摘要

The advent of the machine learning era means increasing data volumes and real-time data processing requirements especially in the medical field to improve the reliability of treatment results. Traditional data indexing has requirements on the storage space of medical devices and does not meet the characteristics of lightweight smart devices. In this paper, we propose a multi-dimensional medical data learned index, design novel mapping function for multi-dimensional data processing, and index multi-dimensional data using segmented linear function. It can efficiently perform data access and indexing processing on lightweight intelligent data processing devices and transmit abnormal data to medical centres. After a series of experiments, the results show that the proposed index supports direct data processing on lightweight intelligent data processing devices, which outperforms the traditional index by more than 10 times.