Using Artificial Intelligence to Predict Decisions of the Turkish Constitutional Court


Sert M. F., Yıldırım E., Haşlak İ.

SOCIAL SCIENCE COMPUTER REVIEW, pp.1, 2021 (SSCI)

  • Publication Type: Article / Article
  • Publication Date: 2021
  • Doi Number: 10.1177/08944393211010398
  • Journal Name: SOCIAL SCIENCE COMPUTER REVIEW
  • Journal Indexes: Social Sciences Citation Index (SSCI), Scopus, Academic Search Premier, FRANCIS, IBZ Online, Aerospace Database, Communication Abstracts, EBSCO Education Source, Educational research abstracts (ERA), Index Islamicus, INSPEC, Library, Information Science & Technology Abstracts (LISTA), Metadex, Political Science Complete, Psycinfo, Public Affairs Index, Social services abstracts, Sociological abstracts, Worldwide Political Science Abstracts, Civil Engineering Abstracts
  • Page Numbers: pp.1
  • Recep Tayyip Erdoğan University Affiliated: Yes

Abstract

This article draws on an artificial intelligence (AI) technique to predict whether an individual application regarding the Turkish Constitutional Court’s public morality and freedom of expression cases leading to a “violation” or a “nonviolation” decision. To this end, four different data sets have been composed, preclassification and fundamental word embeddings steps have been made on each data set. Multilayer perceptron, which is based on artificial neural networks, has been used for the prediction of the case decisions. We have predicted the court’s decisions on these cases with the high success rates (average accuracy of 90%) by using the subject or reasoning sections of texts of the cases as data. The subject section of the cases constituting only a very small part of the data has yielded the highest accuracy. The article has demonstrated that a basic AI technique can be successful in achieving accurate predictions even with relatively small data sets derived from well-structured court rulings.