With the development of internet shopping, the amount of user data generated is increasing day by day. In this paper, a shopping recommendation system based on deep learning is constructed. The user data crawling module and shopping recommendation module are mainly designed. Firstly, obtain important user review information and product information from Jingdong Mall by python crawler and build a user data crawling module. Then a shopping recommendation system was constructed based on deep learning, combined with recommendation algorithm, The system extracts the characteristics of users and commodities through neural network algorithms, proposing a coupled recommendation algorithm (referred to as U-S recommendation algorithm) based on user characteristics and product similarity. The algorithm calculates the best match rate between users and commodities. The results show that the proposed algorithm can improve the effectiveness of the recommendation system, compared with the algorithm based on similarity of products.
In mobile peer to peer streaming system, to solve the influence brought by the load difference among cells, a neighbor peer selection algorithm based on load balance is proposed. In the strategy, the comprehensive evaluation value of the candidate peer is achieved through the use of gray relation analysis theory on multi-attribute decision matrix including load balance degree, and then the neighbor peer list is got through ranking the comprehensive evaluation value. Simulation results show that with the peer selection scheme in mobile p2p streaming system, the system's bandwidth utilization and the user's average throughput are improved.
To combat the challenges(the mismatch between throughput and layer delivery ratio, the useless packets in high layers due to the loss of packets in low layers) brought by the layered encoding to data scheduling for P2P streaming in the mobile network, We propose a new data scheduling scheme, named MLayerP2P. The key idea of the MLayerP2P is: A 3-Stage Model is used to schedule the absent streaming data, where the missed data with high importance is requested in high priorities and a variable windows number is determined in the Free Stage, in the Decision stage the first step of the subscribed layers decision is made, and the second step is made in the Remedy Stage, in the Remedy Stage the available bandwidth is firstly allocated to the missed blocks in the base layer. Simulation results show that the MLayerP2P outperforms other schemes in throughput, delivery ratio, useless packet ratio and interrupt times obviously.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.