A routine for parameter optimization using an accelerated grid-search method

https://doi.org/10.1016/0010-4655(87)90042-7Get rights and content

First page preview

First page preview
Click to open first page preview

References (6)

  • Z. Basrak

    Comput. Phys. Commun.

    (1987)
  • Z. Basrak et al.

    Nucl. Phys.

    (1985)
  • J. Lang et al.

    Comput. Phys. Commun.

    (1971)
There are more references available in the full text version of this article.

Cited by (11)

  • Hyper-parameter optimization of multiple machine learning algorithms for molecular property prediction using hyperopt library

    2022, Chinese Journal of Chemical Engineering
    Citation Excerpt :

    However, the published studies either employed arbitrarily selected hyper-parameters [13] or explored the hyper-parameters through the grid searching method [13,14]. The grid search method works well in searching the entire search space and the global optimal solution can be achieved when the interval of grid search is employed small enough [15], while a huge amount of computing resources will be consumed. Typically, machine learning models training on different datasets correspond to different optimal hyper-parameters [16], and therefore there would be a very small probability of empirically selecting the optimal hyper-parameters as shown in Fig. 1(a), (c), and (d).

  • Idea mining for web-based weak signal detection

    2015, Futures
    Citation Excerpt :

    The reduced number of parameters enables to calculate an optimized value for each of the selected parameters in order to get the highest cross-validated F-measure (Guns, Lioma, & Larsen, 2012). This is realized by applying a grid search (Basrak, 1987; Jiménez, Lázaro, & Dorronsoro, 2009) where discrete sequences are used to set the parameter values. The proposed approach is depicted in Fig. 1.

  • Web mining based extraction of problem solution ideas

    2013, Expert Systems with Applications
    Citation Excerpt :

    This enables calculating an optimized value for each of the selected parameters without high computationally costs. A grid search (Basrak, 1987) is applied where discrete sequences for the parameter values are used. These values are selected that results in the highest cross-validated F-measure (Guns, Lioma, & Larsen, 2012).

View all citing articles on Scopus
View full text