Identification of the use of unauthorized apps in the o2o service by combining online events and offline conditions

Changoh Kim, Huy Kang Kim

Research output: Contribution to journalArticlepeer-review

Abstract

A model for detecting unauthorized Apps use events by combined analysis of situation information in an offline service and user behavior in an online environment is proposed. The detection and response to abnormal behavior in the O2O service environment can be focused on providers, whose decisions change dynamically based on the offline market status and conditions. However, the method for identifying the user’s tools and detecting the usage pattern of the service user were developed in the existing online service environment. Thus, in order to identify abnormal behavior in the O2O service environment, we conducted an experiment to identify the abnormal behavior of providers of smart mobility services, a representative O2O service. In the experiment, the range of normal behavior of a taxi drivers was identified, which was prepared on the basis of the test result directly executed by an expert. The optimal features were selected in order to effectively detect abnormal behavior from the event data relating to the service call acceptance behavior. In addition, by processing the collected data based on the selected features by using various machine-learning classification algorithms, we derived a detection and prediction model that is 98.28% accurate with a prediction result of more than 74% based on the F1 score. Based on these results, we expect to be able to respond to abnormal behavior that may occur in various types of O2O services.

Original languageEnglish
Article number1977
Pages (from-to)1-25
Number of pages25
JournalElectronics (Switzerland)
Volume9
Issue number11
DOIs
Publication statusPublished - 2020 Nov

Keywords

  • Abnormal behavior detection
  • O2O service
  • Unauthorized method

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Signal Processing
  • Hardware and Architecture
  • Computer Networks and Communications
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Identification of the use of unauthorized apps in the o2o service by combining online events and offline conditions'. Together they form a unique fingerprint.

Cite this