Research On Service Recommendation In Social Network Environments

With the rapid increase in the use of web services having similar functionalities in our daily activities,there has been an exponential increase in information needed to aid in making an informed decision.Consequently,it is very challenging to select a superior service that suits the requirements of the service user from a large list of functionally equivalent services.Service recommendations in social network environments have thus become an important issue in the area of web services.In response to increasing numbers of services in the recommendation process,there is a corresponding increase in service consumers.Increases in both sides lead to a diversity in the demand and supply of services.Furthermore,most of the traditional approaches recommend services based on item ratings by users and suffer the issues of data sparsity.To this end,incorporating social networks in service recommendations can help to recommend services with reliability and accuracy along with providing diverse options to users.In this dissertation,first,we proposed a neighborhood-based diffusion recommendation algorithm(NBD)that distributes the resources among items using the rating scores of the items based on the preferences of the target user neighbors.This is done to ensure that the items recommended are personalized to the target user.Specifically,the Adamic-Adar similarity index is explored to select the top-K similar neighbors during the first stage of the diffusion process to suppress popular neighbors.A sigmoid function is then applied to redistribute resources back to the services at the second stage of the diffusion process to suppress popular services.Several experiments have been conducted using four real-world datasets(Friendfeed,Epinions,Movie Lens-100 K,and Netflix)by comparing with some baseline methods.The results show that our proposed NBD algorithm performs better in terms of accuracy and second to the Heat diffusion method in terms of diversity and novelty.The second aspect of the work in the dissertation is the process of friendship-based web service recommendation.We proposed a new approach for item recommendation based on a diffusion method that combines user relationships in social networks with a user-item relationship.Especially,a resource redistribution process is explored in the user–service network that gives mass diffusion a higher recommendation accuracy and Heat Conduct a greater diversity by considering the social degree of users whilst calculating the user degree in the network.A tuning parameter is introduced to adjust the weight of resources that the services finally receives from users based on their social relationships.Extensive experiments conducted on the real-world datasets(Friendfeed and Epinions)which contain friendship relationships,demonstrate the efficiency of our proposed method in achieving notable performance improvements in terms of the recommendation accuracy,service diversity,novelty,and practical dependability.The third part of the dissertation proposes an effective neighborhood discovery algorithm based on local information using link prediction.There has been much interest in link prediction research with significant studies on how to predict missing links or future links in a network based on observed information.However,the key solution to tackle the link prediction problem is how to measure the similarity between the nodes in a network with higher accuracy.Several techniques have been proposed that utilize the similarity between nodes to estimate their proximity in the network.In this dissertation,an efficient link prediction algorithm that predicts relationships between links using the network structure is proposed.This algorithm uses common neighbors in addition to the degree distribution of the nodes to estimate the possibility of the presence of a link between two nodes in a network based on local information.Extensive experiments are carried out and compared with 12 standard similarity-based methods using 7 real-world datasets.The experimental results show that our method has achieved higher prediction accuracy when compared with the local information based methods like the Common Neighbor and Preferential Attachment.It is also competitive with the quasi-local indicators such as LP and global indicators like Katz,with a lower computational complexity than the two.