- 期刊
CTO ORDER CLUSTERING FOR ELECTRONIC PRODUCTS BASED ON WORD VECTORS
摘要
The business model of configure to order has been developed on a large scale, and orders configured according to user needs are an important way to transform producers. Effective clustering of CTO (Configure To Order) orders can effectively reduce the operational costs of producers. In order to solve the above problem, a clustering model for CTO orders is developed in the analysis of user order configurations and a vector representation based on orders is designed. First, the Skip-gram model was applied to train on the electronic product dataset to obtain word vectors; then the similarity between orders was calculated using the vector representation of orders. The experimental results show that the algorithm has better vector differentiation ability in the similarity measure of orders compared to the topic-based feature representation and the word frequency-based feature representation. The clustering metric shows that the SG-CTO algorithm has better results.
延伸閱讀
- 蘇千傑（2007）。Document clustering based on vector space model with concepts as the dimension value〔碩士論文，國立中央大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0031-0207200917351030
- 石耿維（2009）。Robust clustering for cluster ensemble〔碩士論文，國立交通大學〕。華藝線上圖書館。https://doi.org/10.6842/NCTU.2009.00917
- 林子敬（2018）。Application of User Clustering Based on Topic Modeling〔碩士論文，淡江大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0002-0708201813110800
- 陳垂呈（2019）。Using Clustering Method with Products-Centric to Find Adaptive Consumers。資訊與管理科學，12(2)，19-28。https://www.airitilibrary.com/Article/Detail?DocID=20720971-201912-202002240008-202002240008-19-28
- 毛泰源（2010）。Token-Based Clustering Method in Vehicular Ad-Hoc Network〔碩士論文，國立交通大學〕。華藝線上圖書館。https://doi.org/10.6842/NCTU.2010.00700