Research Of Fashion Conceptual Design Method Based On Fashion Knowledge Management

In the context of innovation-driven reformation and development of fashion industry in China, it becomes the most essential issue to enhance the ability of independent R&D and creative design level for Chinese local fashion brands. In quite a long period, fashion is considered to be determined by fashion designers. However, fashion is hereby considered to be formed according to certain social background, instead of being determined by certain people’s subjective minds. So fashion could be generated by precise analysis from objective factors. Now in the context of fast fashion, fashion design is not merely rely on the designers’ creativity, but all kinds of modern information technology are applied in the process of fashion design. The dissertation is formed by the National211Sub-project "fashion art creativity and trend optimization", based on the idea of knowledge management, using the tool of data warehouse to imitate the process of fashion designers’ subjective fashion optimization and judgment, thus to provide scientific, objective, fast and accurate decision support for fashion conceptual design. The main content of the paper is as following:1. According to the demand of fashion knowledge management, the paper proposes the concept of Fashion Data Warehouse (FDW) based on knowledge management, and analyzes the necessity and feasibility of using data warehouse technology. According to the characteristics of the fashion data warehouse system, an overall structure composed of fashion data dictionary, fashion data sources, fashion data management, fashion data mining and the front-end decision support is formed. The proposed concept of Fashion Data Dictionary (FDD), including Fashion Color Data Dictionary, Fashion Material Data Dictionary, Fashion Accessory Data Dictionary, Fashion Pattern Data Dictionary, Fashion Technique Data Dictionary, Fashion Style Data Dictionary, and Fashion Look Data Dictionary is formed, in order that all kinds of fashion data from different sources are unified in format. Each data dictionary regulates its data type, level, content and standard presentation. Sources of fashion data extraction are fashion clothing, social background, and art works. Fashion clothing data sources include fashion shows, fashion market, fashion brand advertisement, target consumer, fashion e-shop, fashion and fabric exhibition, etc. Social background data sources include politics, economy, environment, science and technology, sports, lifestyle, etc. Art works data sources include TV drama, art, design, music, performance art, literature, etc. The fashion data management is defined including fashion data extraction, naming method, conversion rules, and loading standard, so that the fashion data extracted from a variety of sources could be loaded in the fashion warehouse with standardized data format.2. A multidimensional data mining method for fashion knowledge is discussed. In order to solve the problem of controlling fashion designers’ fashion knowledge, including information sources, quantity, quality, understanding, and analysis level of fashion knowledge, based on fashion analysis demand and OLAP cube data mining principle, the paper discusses the design of fashion data mining dimension table, including the time-dimension, source-dimension, brand-dimension, color-dimension, fabric-dimension, accessory-dimension, pattern-dimension technique-dimension, style-dimension, look-dimension levels and members, and develops an OLAP star-model. According to fashion data multidimensional characteristics, fashion data cube and the cube model of various dimensions of data mining are defined, including fashion color, material, accessory, pattern, technique, style, look, and accordingly the target of fashion analysis, data mining hierarchy, data cube set, data mining result example and validity. Through OLAP multidimensional data mining, the objective knowledge of fashion, including fashion colors, fashion materials, fashion accessories, fashion patterns, fashion techniques, fashion styles and fashion looks could be acquired. As results for data mining, the fashion data proportion statistics, order statistics and key snapshot of front-end reports for presentation could provided with quantitative and visualized conclusions, so as to provide objective references for fashion conceptual design. The paper takes fashion color as example to show the actual effect of OLAP data mining result. Meanwhile, the paper discusses the OLAP data mining methods of fashion data from the social background and art work sources.3. Social background has an important impact on the formation of fashion style is the consensus of the fashion industry, but the study of the relationship between the two has always been to stay in the sociology of qualitative research. The focus of the paper is the methodology of turning the implicit knowledge of social background into the explicit knowledge of fashion style. The fashion style tendencies are concluded into opposite directions:retro-futuristic, luxury-simple, conservative-individual, neutral-gender differences. And the styles are divided into realistic style tendency and ideal style tendency. With the historical research of western fashion and the social background of politics, economy, environment, science and technology events from1900to2009, the regulations between fashion style tendency and social background are concluded. The social background is indexed as "bipolar type" and "monopole type", and the concept of the social background Composite Index is put forward. By extracting the average figure, peculiar figure and the concentrated area of Composite Index, the fashion style tendency could be drawn. The paper demonstrated the feasibility of the conversion from the implicit knowledge of social background to explicit knowledge of fashion style through example research. The advantage of this approach is breaking through time limit and regional restrictions, so that the fashion styles can be drawn from various time period and districts.4. The concept and structure of fashion conceptual design decision support is proposed in the paper. Fashion conceptual design is with creativity, multiple solution, hierarchy, similarity, empiric and comprehensive features, and is the most essential and most creative stage in the design process. The paper presents a conceptual design decision support method based on the fashion data warehouse, in order to avoid blindness and subjectivity in the conceptual design, to achieve objective, accurate and efficient conceptual design results. The definition of fashion conceptual design is the fashion creative concepts for guiding the new fashion product development, including the concept of design theme, and the corresponding color, material, accessory, pattern, technique and style design. The method of fashion conceptual design decision support for fashion themes and elements based on data warehouse is discussed. The paper also plans for the human-computer interaction of fashion conceptual design, including the structure, process and interface design. Decision support results from system can be stored, exported, shared and updated, so that the accumulation and reuse fashion conceptual design knowledge could provide possibility for the real-time collaborative design teamwork. The paper conducted experiments on fashion conceptual design based on decision support, to demonstrate the feasibility of the process of keyword selection, conceptual design object establishment, fashion style theme concept recommendation, theme concept summary interface, and color conceptual design.The innovation of the paper is as the following four areas:1. The concepts of Fashion Data Warehouse and Fashion Data Dictionary are proposed. Fashion data could be converted with the standard of the data dictionary, including fashion color, material, accessory, pattern, technique, style and look. Various fashion data from different sources could be unified in format, thus to meet requirements for subsequent data mining and decision support functions.2. The online analytical processing (OLAP) method of fashion knowledge multidimensional data mining is proposed. The multidimensional model is designed according to the various fashion data. Also the designs of data mining statistical reports are discussed.3. The method of turning the implicit knowledge of social background into the explicit knowledge of fashion style with composite index is proposed. The paper tries to make a shift from qualitative research to quantitative research in fashion style.4. A fashion conceptual design decision support method based on fashion data warehouse is proposed. Through data mining, the recommended fashion style theme and corresponding color, fabric, pattern, technique, style concepts are suggested for conceptual design.The results of the paper will be applied in the fashion enterprises and public service platform, in order to help the companies to enhance the scientific objectivity, accuracy and efficiency in fashion conceptual design, thus to enhance transformation, upgrading, and raise level of independent R&D for fashion industry in China.