| In the era of mobile Internet,intelligent services have gradually become the research focus in the fields of service computing,cloud computing and mobile edge computing,and the active perception of service demand is the key to realize intelligent services.In a dynamic and complex environment,how to dynamically predict user service requirements based on the context information of the user’s current environment is one of the key issues that need to be addressed.Therefore,this study has conducted research on dynamic prediction algorithms for user service requirements based on context awareness,mainly including the following aspects:Aiming at the problems of fuzzy C-means clustering algorithm(FCM),this study proposes a fuzzy C-means clustering algorithm based on multi-strategy flower pollination optimization(WGFFCM).Firstly,the algorithm introduces a chaotic mapping sequence to optimize the initial position of the pollen population,improving the overall convergence accuracy and the quality of the optimal value.Secondly,it uses inertial weighting factors and the golden sine algorithm to improve the convergence accuracy and the ability to jump out of local optimums.Finally,it uses a multi-strategy flower pollination algorithm to optimize the selection of the initial cluster center of the FCM algorithm,solving the sensitivity of the FCM algorithm to the initial cluster center,it is easy to fall into problems such as local optimization to achieve better clustering results.Regarding the sparsity of service data types for individual users,this study extracts user service rating information and service item feature information from the service dataset to construct a user preference matrix.Based on the user preference matrix,the WGFFCM algorithm is used to cluster similar users into the same cluster,greatly alleviating data sparsity,reducing subsequent prediction model training time,and improving the accuracy of service demand prediction.This study proposes a dynamic prediction method for service demand based on context awareness hidden Markov model(HMMCA)to address issues such as changing user preferences and different user needs in different context.Firstly,set a similarity threshold based on clustering users δ,find the nearest neighbor user set NS(u)of the user cluster class.Secondly,use the service data of the nearest neighbor set NS(u)as the initial parameter of the hidden Markov model(HMM)to simulate user preference transfer to dynamically predict user service requirements.Finally,context similarity is calculated based on the post context filtering(CA)method for users in the current context.Attributes that affect user preferences under a single context information are adjusted based on the weight of the user’s service needs under different context information,thereby recommending them to users. |
