信息化-电子商务-电子商务背景下的女装搭配推荐系统研究.pdf

The booming development of Internet technology has made online consumption

become the mainstream of daily consumption, and it has also brought e-commerce to

solve the "information overload" problem after information retrieval technology.

However, with the development of social productivity and economy, consumers'

consumption concepts are constantly changing from the satisfaction of the basic

necessities of life to the pursuit of psychological needs. Traditional commodity

recommendation information has even been considered as "Spam bombing". The

development of recommendation technology needs to start from the psychological

needs of consumers.

This study takes the dress collocation recommendation as the research direction,

summarizes and contrasts the recommendation of related platforms, takes brand L as

the research object, combines the brand design elements and literature research to

The main contents of the study are as follows:

Three typical shopping websites, two typical dress collocation websites, some

clothing brand official websites and official apps, and three types of dress collocation

apps were selected, their recommendation principles, service providers, push locations,

recommendation quantities, restrictions and purchase process were summarized and

compared. The shortcomings of those platforms in the aspect of dress collocation

recommendation service were discussed, which provided reference for how to

which could characterize brand L's fashion style best were finally selected. Using the

semantic differentials method to construct a fashion style kansei evaluation scale, nine

people who have professional background conducted evaluation experiments on 1000

brand L's clothing pictures.

characteristics, the design elements of the clothing were subdivided into three aspects:

style, color and fabric, and the classification standard was formulated, totaling 10

theory, the 1000 pictures were encoded to prepare for subsequent data analysis.

established , and fashion style matching rules followed the general matching principle.

Based on the fashion style quantification model, the similarity and other algorithms

were used to establish the women's dress collocation recommendation system. The

relevant evaluation indicators and user subjective evaluation were used to verify the

recommendation results of the system. Finally, the characteristics and application

scenarios of the system were summarized.

recommendation system