以美感生活型態、藝術偏好與人格特質探討網路服飾之推薦機制

Using the Aesthetics Lifestyle, Artistic Preference and Personality trait to Explore the Recommendation Mechanism of Man’s Apparel on Line

葉志發(JR-FA Ye)

指導教授 ： 陳瑾儀

中原大學/商學院/企業管理研究所/碩士(2012年)

https://doi.org/10.6840/cycu201200369

摘要

摘要
　　近年人們生活水準大幅提升，且現今網路上的商品種類眾多，目前的網路服飾推薦仍以熱門或折扣推薦為主，對於個人化推薦尚未普及，為幫助消費者做正確選擇，因此希望推薦機制與網路服飾結合，以男士服裝為出發點，發展出可以推薦消費者合適衣物的網路服飾推薦機制。
　　在探討推薦機制的變數上，本研究以美感生活型態及藝術偏好探討男性服飾風格，以決策樹發展分類機制，並以二階段實驗問卷驗證推薦滿意度，並分組探討示範效果對於不同人格特質(高低自我監控)之影響。
　　第一階段問卷樣本回收後，經分類迴歸樹 (CART)分類服飾風格，服飾風格分為：搖滾風格、學院風格、街頭風格、美國風格及嘻哈風格等五類，每類服飾風格皆有其特色服飾單品，和美感生活型態及藝術偏好的特徵，而此階段分類正確率為54.3%。
　　除此之外，本研究又進行第二階段的實驗問卷，驗證後此階段推薦機制滿意度為70%，而實驗共分為A、B、C三組，作關於示範效果的實驗，A組為對照組，B組及C組為示範效果組，B組示範單品棒球外套，C組則示範嘻哈風格的穿著，並以人格特質-自我監控來探討其結果，且結果顯示示範效果對消費者偏好影響是不顯著的，示範效果對於自我監控上，B組棒球外套部分低自我監控者反而會受示範效果影響，C組嘻哈風格部分，低自我監控者則不受示範效果影響，而兩者高自我監控者，皆不受示範效果影響。網路服飾可根據本研究的分類樹推薦對消費者進行交叉銷售，且可在根據各服飾風格的特色，發展設計自身的產品，減少錯誤的推薦，使購買產品的意願增加，提升網路服飾店的銷售。

關鍵字：男性服飾風格、美感生活型態、藝術偏好、分類迴歸樹 (CART)、網路服飾推薦、人格特質

關鍵字

人格特質 ； 男性服飾風格 ； 美感生活型態 ； 網路服飾推薦 ； 分類迴歸樹 (CART) ； 藝術偏好

並列摘要

Abstract
　　Standard of living increases dramatically in recent years, and products can be bought easily online. To help consumers do their right choices, the concept of recommendation system is developed these days. Until now, the recommadation systems online still focus on popular or discount items. The contribution of this thesis is to combine the concept of the recommendation system and personal apparel items to assist consumers choose proper merchandises for themselves. This thesis starts from men's clothing, aims to explore a recommadation mechanism to help consumers choose appropriate clothes.
Variables applied to discuss man's apparel style in this thesis are aesthetics lifestyle and artistic preference. Furthermore, this study uses two-stage experimental questionnaires to identify the satisfaction toward the recommendation mechanism developed in the first stage. Also, subjects in the second stage are separated into three different groups to discuss the demostration effect to different personality traits (high and low self-monitoring).
After the data processing, the dress styles can be classified by classification and regression tree (CART) into Rock Style, College Style, Street Style, American Style and Hip-hop Style. Each style has specific features with different characteristics of the aesthetics lifestyle and artistic preference, in this phase, the correct classification rate of the decision tree is 54.3%.
In the second phase of the experiment, instead of verifying the satisfaction of recommendation mechanisms is 70%. Subjects were divided into A, B, C groups to explore demonstration effect. Group B and group C were the experimantal groups while group A is the comparison group. In the experiment, a powerpoint and statement of a single product (baseball jacket) was shown to B group, and the hip-hop style of wearing was demonstrated to group C. Then personality traits - self-monitoring was applied to investigate the results. In the end, the result shows non-significant in the demonstration effect of consumer preferences. On the other hand, group B shows low self-monitoring subjects were influenced by the demonstration effect while group C shows low self-monitoring subjects were not influenced by the demonstration effect, and both high self-monitoring subjects in group B and C shows no impact on demonstration effect. Also, using the classification tree, cross-selling of apparel online can be possible, and stores selling apparel online can design their own products in accordance with the characteristics of each style, to reduce the error recommendation, increase willingness to purchase and enhance the sale of online apparel stores.
Keywords ：Man’s apparel style、Aesthetics Lifestyle、Artistic Preference、Classification and regression tree (CART)、Recommendation of Apparel on Line、Personality Trait

並列關鍵字

Personality Trait ； Classification and regression tree (CART) ； Aesthetics Lifestyle ； Recommendation of Apparel on Line ； Man’s apparel style ； Artistic Preference

參考文獻

吳肇銘（2008），以消費者購買決策階段為基礎之適性化推薦系統設計與績效之研究，資訊管理展望， 10(2), 19-43。

許士軍（1986），“管理學”，東華書局。

陳明陽（2010），美感生活型態對產品偏好的影響，交通大學應用藝術研究所碩士論文。

蘇宛蓁（2008），品牌忠誠度與從眾行為之關係探討，中原大學商業設計學系碩士論文。

吳肇銘（2004），以消費者購買決策為基礎之適性化推薦系統，中原大學資訊管理學系碩士論文。