Tea Product Recommendation Based On The Time Context Of Implicit Semantic Model

With the rapid development of e-commerce of agricultural products,tea products e-commerce platform is also developing rapidly.Tea is drink whose consumption is the second largest in the world,tea and people's lives are closely related.According to the data which show that,in 2013,the total sales of tea through the electricity business channel close to 8.5 billion Yuan,nearly three times more than in 2012,the rapid growth;in 2014,its totaled sales is 11.3 billion Yuan that exceeded 10 billion Yuan output at the first time,an increase of 32.94%compared with last year;2015 online sales market is nearly 12 billion,maintaining high growth.With the rapid development of tea products in e-commerce,How to find tea products which interest user in a lot of information become a hot topic of research.As a personalized recommendation technology to solve the problem of information overload,the current research and application in e-commerce such as books and movies is becoming more and more extensive,and it will be the inevitable trend of future development of tea products e-commerce.In this paper,the tea product e-commerce as the research object,taking into account the user(consumer)interest with the impact of time changes,mainly based on the implicit semantic model(Latent factor model,referred to as LFM)time context of tea products personalized recommendation.The main work of this paper is as follows:(1)Aiming at the shortcomings of the existing latent semantic model LFM in solving the cold start problem,this paper proposes a recommendation algorithm based on user attribute latent model-Attributes-LFM.The user's attribute(age,gender,etc.)information is used to find the neighborhood user whose attributes are similar to the user,and then use the similar neighborhood user's interest preference to refer to the current user.Experiments show that the new algorithm effectively alleviates the cold start problem.(2)According to the characteristics of user's interest in the existing recommendation system,this paper introduces the context information of the time of the tea product into the Attributes-LFM algorithm on the basis of(1)A Time Context Recommendation Algorithm Based on Implicit Semantic Model-TAttributes-LFM.The new algorithm is based on the traditional "user-project" two-dimensional recommendation process,and constructs the "user-project-time" three-dimensional matrix.By changing the matrix decomposition form of the model,the decomposition of the "user-project-time" three-dimensional scoring matrix is realized.Improve the accuracy of the model at the same time to achieve the expansion of the model.(3)TAttributes-LFM type verification and evaluation.The TAttributes-LFM algorithm was validated by using the data of Jingdong Mall.Experiments show that the new algorithm effectively improves the accuracy of the proposed results and alleviates the cold start problem.(4)Finally,design and implementation of a TAttributes-LFM based on tea products recommended prototype system.The research of this paper has a certain impetus to promote the future development of e-commerce of tea products.