Scholarlink

<p>Feature selection is an essential pre-processing step to lower dataset dimensionality for the most useful feature optimization in big data applications. As a result, we presented a unique stochastic optimization technique in this research that uses a hybrid multi-objective optimization strategy for feature selection&#xa0;in big data optimization problems. This method combines the Osprey optimization algorithm (OSA) with This algorithm is called differential evolution (DE). The differences Evolution increases the information exploration capabilities of the Osprey Optimization method by utilizing its operators as local search mechanisms. Overall, the suggested approach consists of three steps. First, Both the population and the archive are created. The hybrid OSA and the DE algorithm are used in the second stage to update. The last stage is finding the solutions and nondominated solutions and maintaining the collection. The effectiveness of the recommended method is assessed using standard deviation, average, and best metrics, and the results are compared to those of alternative algorithms. Our experiment findings showed that the suggested strategy performed better than alternative strategies.</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

Multi-objective hybrid stochastic scheme for big data optimization problem using osprey optimization algorithm and differential evolution

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • E. Ramesh Babu,
  • M. Sunil Kumar

摘要

Feature selection is an essential pre-processing step to lower dataset dimensionality for the most useful feature optimization in big data applications. As a result, we presented a unique stochastic optimization technique in this research that uses a hybrid multi-objective optimization strategy for feature selection in big data optimization problems. This method combines the Osprey optimization algorithm (OSA) with This algorithm is called differential evolution (DE). The differences Evolution increases the information exploration capabilities of the Osprey Optimization method by utilizing its operators as local search mechanisms. Overall, the suggested approach consists of three steps. First, Both the population and the archive are created. The hybrid OSA and the DE algorithm are used in the second stage to update. The last stage is finding the solutions and nondominated solutions and maintaining the collection. The effectiveness of the recommended method is assessed using standard deviation, average, and best metrics, and the results are compared to those of alternative algorithms. Our experiment findings showed that the suggested strategy performed better than alternative strategies.