Data analysis in health and big data: A machine learning medical diagnosis model based on patients’ complaints

SİLAHTAROĞLU G., Yılmaztürk N.

Communications in Statistics - Theory and Methods, vol.50, no.7, pp.1547-1556, 2021 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 50 Issue: 7
  • Publication Date: 2021
  • Doi Number: 10.1080/03610926.2019.1622728
  • Journal Name: Communications in Statistics - Theory and Methods
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Business Source Elite, Business Source Premier, CAB Abstracts, Compendex, Veterinary Science Database, zbMATH, Civil Engineering Abstracts
  • Page Numbers: pp.1547-1556
  • Keywords: Machine learning, text mining, patients' complaints, diagnosis, big data
  • Istanbul Medipol University Affiliated: Yes


The emergence of big data made it possible to make better predictions and discover hidden patterns which contain a load of useful information. Like other domains, health discipline is also enjoying this new data science era. In this study, we suggest some big data sources for health sector, draw a big data framework in health, and we present a machine learning pre-diagnosis model for emergency departments. The system predicts the diagnosis with a minimum accuracy of 75.5%. Patients’ verbal complaints about their own situations are used for machine learning. Two different models, Probabilistic Neural Network based on the Dynamic Decay Adjustment and Random Forest Decision Tree, have been used for machine learning after a series of text mining processes. Although there are other studies to predict diagnosis, this study is probably the first one using patients’ natural verbal complaints as user generated data. Both models’ accuracy and precision statistics suggest that they can be used as a decision support system to direct emergency department patients to appropriate healthcare centers. The system may also be developed into a triage prediction model at emergency departments.