PRESENTING A DECISION TREE-BASED DATA MINING MODEL TO INCREASE THE SECURITY OF USERS' INFORMATION IN THE USE OF ELECTRONIC SERVICES CONSIDERING INTERNET BANKING SYSTEM
PRESENTING A DECISION TREE-BASED DATA MINING MODEL TO INCREASE THE SECURITY OF USERS' INFORMATION IN THE USE OF ELECTRONIC SERVICES CONSIDERING INTERNET BANKING SYSTEM

Summary/Abstract: THE MAIN GOAL OF THIS RESEARCH STUDY IS TO PRESENT A DATA MINING MODEL BASED ON A DECISION TREE TO INCREASE THE SECURITY OF USERS' INFORMATION IN THE USE OF ELECTRONIC SERVICES. TO DETERMINE THE SECURITY OF USERS' INFORMATION IN THE USE OF ELECTRONIC SERVICES. THEREFORE, BASED ON THE DATA AVAILABLE IN THE DATABASE OF 4376 CUSTOMERS OF ELECTRONIC SERVICES IN THE BANK OF AGRICULTURE, THE DATA WAS ANALYZED IN THE DATA MINING MODELS OF CLASSICAL DECISION TREE, NAIVE BIZ, RANDOM FOREST, ROOT TREE, DEEP LEARNING AND NEURAL NETWORK. THEREFORE, BY RUNNING THE RAPIDMINER SOFTWARE, THE OUTPUT OF THE MODEL WAS ANALYZED IN THE PARAMETERS OF ACCURACY, CLASSIFICATION ERROR PERCENTAGE, KAPPA COEFFICIENT AND ABSOLUTE ERROR, AND THE RESULTS SHOWED THAT THE CLASSIC DECISION TREE MODEL AND THE RANDOM FOREST MODEL WERE THE BEST WITH 99.42 PERCENT EACH. HAVE WHILE THE ROOT TREE IS NOT IN A GOOD CONDITION WITH 87.88, IN RELATION TO THE CLASSIFICATION ERROR, THE CONDITIONS OF THE CLASSIC DECISION TREE AND DECISION FOREST MODELS AND THEN THE NEURAL NETWORK MODEL WERE FOUND TO BE MORE SUITABLE THAN OTHER DATA MINING MODELS; IN RELATION TO THE KAPPA ERROR, THE CLASSICAL DECISION TREE AND RANDOM FOREST OBTAINED A MORE FAVORABLE SITUATION THAN OTHER METHODS WITH 0.989, ALSO IN RELATION TO THE ABSOLUTE NORMAL ERROR, THE CLASSICAL DECISION TREE AND RANDOM FOREST EACH SHOWED A LOWER ERROR WITH A RATE OF 0.039.

• Details
• Contents