Kyonggi University

About: Kyonggi University is a education organization based out in Suwon, South Korea. It is known for research contribution in the topics: Catalysis & Dielectric. The organization has 1946 authors who have published 4404 publications receiving 64791 citations.

A new approach to travel information sources and travel behaviour based on cognitive dissonance theory

Abstract: This study examines the influences of en route information on travel behaviours, based on cognitive dissonance theory, which explains that selective exposure to new information leads to the presence of dissonance. Fifteen travel activities were identified from the related literature to measure levels of unplanned travel behaviours. Moreover, two types of information sources used during trips were tested to evaluate the degree of dissonance in accordance with new information provided to travellers while on vacation. The results indicated that during trips, those who use information technology change their intended behaviours, while those who use ‘traditional’ information sources actualize their intended behaviours. In line with cognitive dissonance theory, information technology is viewed as dissonance-increasing information, while traditional information sources used during trips are considered consonance-increasing information.

Using Bayesian networks to analyze medical data

Abstract: Due to many possible causes involved with infertility, it is often difficult for medical doctors to diagnose the exact cause of the problem and to decide the correct therapy. A Bayesian network, in general, is widely accepted as an effective graphical model for analyzing biomedical data to determine associations among variables and to make probabilistic predictions of the expected values of hidden variables. This paper presents Bayesian network-based analysis of infertility patient data, which have been collected from the IVF clinic in a general hospital for two years. Through learning Bayesian networks from the clinical data, we identify the significant factors and their dependence relationships in determining the pregnancy of an infertility patient we classify the patient data into two classes (pregnant and not-pregnant) using the learned Bayesian network classifiers. From this medical data mining, we discovered the new domain knowledge that the age of female partner and stimulants like hCG, FSH, LH, Clomiphene, Parlodel and GnRH play the key role in pregnancy of an infertility patient. Through the experiments for investigating the prediction accuracy, Bayesian network classifiers showed the higher accuracy than non-Bayesian classifiers such as the decision tree and k-NN classifier.