Research And Realization On Garment Style Digitalization And Style Rule Mining

Personal Garment Pattern System is a garment pattern recommendation system.It is user-centric.The system studies individual user's needs from his or her own taste of garment pattern style,and recommends proper garment pattern.It is closer to the user needs and more in line with the trend of our time's development than those that study from the perspective of market or product.My study is based on the design and realization of Personal Garment Pattern System. As a main module of Personal Garment Pattern System,the rule mining module generates typical style rules by analyzing the structure and features of typical garment pattern database and utilizing the knowledge of data mining.The realization of this module is the foundation of Personal Garment Pattern System realization.The main contribution of this thesis is as follows:Base on the background of Personal Garment Pattern System,and the needs of system design and realization,garment styles are classified.In this thesis,we mainly focus on eight mature styles of gar nent pattern,i.e.classic style,trendy style,sports style,casual style,elegant style,neutral style,light style,ethical style.As the structure and the parts of garment pattern are the main aspects of the style of garment pattern,this thesis sums up the factors of garment style on the basic of analyzing garment pattern design,and chooses the factors which have great influence on garment style to digitize.Considering the subjectivity of human decision-making process,we introduce decision field theory which can simulate human decision-making process of the degree of clothing style.The results are in line with the assessment given by clothing professionals.A typical style rules base is the foundation of Personal Garment Pattern System.In order to obtain typical style rules which Personal Garment Pattern System needs,this thesis expatiates on the history of data mining development,introduces general data mining process and summarizes data mining methods and technique.Based on full study the algorithm of association rules mining, a betterment algorithm which can fit the style model of garment pattern better and be realized easier is acopted to increase efficiency.We mine the typical garment pattern database and obtain typical style rules finally.After picking and choosing useful rules,typical style rules base is build.