The hidden pattern of data mining technique’s application and performance in health care research

Today, due to the application of information systems in the field of health and the increasing production of electronic data in this field, the efficient and effective use of this data has received much attention. Data mining is one of these methods that is well-known for extracting knowledge from health data and numerous researches in the field of health data mining are done daily. This multiplicity of applications and diversity of data mining techniques have created a challenge in selecting the appropriate technique(s) in different areas of health. Therefore, in this study, with a systematic literature review and data mining methods, the hidden pattern in the application and performance of these techniques has been extracted. The adjusted data set with ۱۰۱ papers and ۱۶ attributes was collected. During the modeling and evaluation phases, six modeling techniques (FP-Growth, Decision Tree (J۴۸), Naïve Bayes, K-Nearest Neighbor, Gradient Boosted Tree, and Bayesian Boosting) were applied in Rapid miner ۹.۸. The results of this study can be used as a road map to select superior techniques in medical data preparation and mining and reduce data analysis time and cost by preventing technique selection errors.

کلیدواژه ها:

Data mining techniques ، Gradient Boosted Algorithm (GBA) ، Health care