DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kim, Hye-jin | - |
dc.contributor.advisor | 김혜진 | - |
dc.contributor.author | Kim, Youngjun | - |
dc.date.accessioned | 2023-06-22T19:31:01Z | - |
dc.date.available | 2023-06-22T19:31:01Z | - |
dc.date.issued | 2022 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=997150&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/308150 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 기술경영학부, 2022.2,[iii, 35 p. :] | - |
dc.description.abstract | A product’s package is the first product element to be exposed to consumers and plays an important role in delivering information. However, prior research has been limited to human judgment and lacks systematic analysis of package elements. This study aims to systematically and quantitatively analyze the package elements used in front package design and identify typologies used in managerial practices. We extracted the features of the package elements by leveraging computer vision techniques and using real-life product package examples. Based on prior literature, we categorized all text and image format package elements. The features of the package elements were size, location, and color. The clustering algorithm revealed that four clusters are the optimum number for the data, with distinctive features: “Kids Appeal,” “Healthy,” “Classics,” and “Niche.” Finally, guidelines with specific numbers are suggested, to assist marketing managers in selecting the right package design elements to achieve the desired positioning. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.title | Analysis of product package design elements through machine learning | - |
dc.title.alternative | 머신러닝을 통한 상품 패키지 디자인 구성요소에 대한 연구 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :기술경영학부, | - |
dc.contributor.alternativeauthor | 김영준 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.