earticle

영어

In 4th Industrial Revolution, the types of data collected are diverse and the scale is growing large. With the popularization of smart devices and the spread of social network services(SNS), lots of information such as items purchased by people, video contents viewed, shopping list, location that we visited and keywords that we searched have been collecting and overflowing countless times. In the industry, all this information is converted to data. Also, various types of information from customers are collected and many companies build and research numerous business models. These days, researches on personalized recommendation services are also being actively conducted, and the form of data collected accordingly is also very extensive. The purpose of this study is to analyze and utilize items of categorical variables to create similar clusters, and to apply association analysis within each cluster to provide meaningful information to people. In this study, after dividing the attribute data sets of Vogue fashion show image without class into top, bottom and whole, and was applied for Spherical K-means clustering and association rules. As a result of the simulation, it was confirmed that a relatively large number of various combinations appeared on the whole, the combination of colors and other properties mainly appeared in the bottom, and we found that the association density was lower than that of the whole and top. In the future, it is necessary to consider additional analysis and in-depth interpretation of what the results of this study mean in the industry and how the information can be used.