Item request has been placed! ×
Item request cannot be made. ×
loading  Processing Request

SVM Optimization with Grid Search Cross Validation for Improving Accuracy of Schizophrenia Classification Based on EEG Signal

Item request has been placed! ×
Item request cannot be made. ×
loading   Processing Request
  • معلومة اضافية
    • بيانات النشر:
      Department of Informatics, Universitas Islam Negeri Syarif Hidayatullah
    • الموضوع:
      2024
    • Collection:
      E-Journal Universitas Islam Negeri (UIN) Syarif Hidayatullah Jakarta
    • نبذة مختصرة :
      The advantage of the Support Vector Machine (SVM) is that it can solve classification and regression problems both linearly and non-linearly. SVM also has high accuracy and a relatively low error rate. However, SVM also has weaknesses, namely the difficulty of determining optimal parameter values, even though setting exact parameter values affects the accuracy of SVM classification. Therefore, to overcome the weaknesses of SVM, optimizing and finding optimal parameter values is necessary. The aim of this research is SVM optimization to find optimal parameter values using the Grid Search Cross-Validation method to increase accuracy in schizophrenia classification. Experiments show that optimization parameters always find a nearly optimal combination of parameters within a specific range. The results of this study show that the level of accuracy obtained by SVM with the grid search cross-validation method in the schizophrenia classification increased by 9.5% with the best parameters, namely C = 1000, gamma = scale, and kernel = RBF, the best parameters were applied to the SVM algorithm and obtained an accuracy of 99.75%, previously without optimizing the accuracy reached 90.25%. The optimal parameters of the SVM obtained by the grid search cross-validation method with a high degree of accuracy can be used as a model to overcome the classification of schizophrenia.
    • File Description:
      application/pdf
    • Relation:
      https://journal.uinjkt.ac.id/index.php/ti/article/view/37422/pdf; https://journal.uinjkt.ac.id/index.php/ti/article/downloadSuppFile/37422/10099; https://journal.uinjkt.ac.id/index.php/ti/article/view/37422
    • الرقم المعرف:
      10.15408/jti.v17i1.37422
    • الدخول الالكتروني :
      https://journal.uinjkt.ac.id/index.php/ti/article/view/37422
      https://doi.org/10.15408/jti.v17i1.37422
    • Rights:
      Copyright (c) 2024 Masdar Desiawan, Achmad Solichin ; https://creativecommons.org/licenses/by-sa/4.0
    • الرقم المعرف:
      edsbas.81D67DE8