Abstract The rapidly increasing data volume produced daily is encouraging to generate novel procedures for extracting suitable information from such data. Machine learning is an application of artificial intelligence which is employed to provide relevant knowledge extracted from data, due to such characteristics, the adoption of machine learning approaches is one of the most accepted alternatives for this purpose nowadays. On the other hand, many machine learning applications turn into complex tasks due to the nature of data and the procedure that these must be subjected to collecting appropriate information. Alternatively, metaheuristic techniques are optimization algorithms widely used for treating complex tasks efficiently. The Gravitational Search Algorithm (GSA) is an optimization method based on the Newtonian gravitational laws and the interaction of masses, this procedure has proved interesting results due to the employed operators for correct balancing the exploration and exploitation stages, avoiding the common flaws present in existing optimization techniques such as the premature convergence onto local minimal. In this study, a comprehensive overview of the GSA applied in several machine learning applications is carried out.

中文翻译：

---

用于机器学习问题的 GSA：全面概述

摘要 每天产生的快速增长的数据量令人鼓舞地产生用于从此类数据中提取合适信息的新程序。机器学习是人工智能的一种应用，用于提供从数据中提取的相关知识，由于这些特性，采用机器学习方法是当今最被接受的替代方法之一。另一方面，由于数据的性质和收集适当信息的程序，许多机器学习应用程序变成了复杂的任务。或者，元启发式技术是广泛用于有效处理复杂任务的优化算法。引力搜索算法 (GSA) 是一种基于牛顿引力定律和质量相互作用的优化方法，由于采用了算子正确平衡了勘探和开发阶段，避免了现有技术存在的常见缺陷，因此该过程证明了有趣的结果。优化技术，例如过早收敛到局部极小值。在本研究中，对 GSA 在多个机器学习应用中的应用进行了全面概述。