Mechanism design for data reliability improvement through network-based reasoning model

Abstract

The importance of business intelligence is increasing every day and many corporations are using data analysis and its results in the decision-making process. Although collecting reliable data is a key prerequisite for the effectiveness of business intelligence, the existing research has focused on improving the reliability of data that is already collected. This research points out the limitations of previous research related to data reliability and presents a new theoretical model that can secure highly reliable data to enhance the business intelligence effect. We designed a mechanism that introduces the concept of the power of influence by using the network effect based reasoning model and applying the two-step flow theory of social exchange theory and information. More than 2 million pieces of real users' preference data was collected and verified by applying them to a recommendation system. More specifically, I made them recognize what the influence of the users' preference data input behavior on each individual would be. Also, the limitation on the user input data with the low existing reliability was overcome by applying the recommendation system based on the network effect, which weights the preference data of the users having high influence. As a result, we have shown that the data collection mechanisms based on influence are more efficient in terms of data collection and data analysis.