In view of the problems that fake reviews mislead users and make their interests suffer losses and the cost of large-scale manual labeling reviews is too high, by using the classification model generated in the previous iteration process to improve the accuracy of detection, a fake review detection model based on Vertical Ensemble Tri-Training (VETT) was proposed. In the model, the user behavior characteristics were combined as features based on the review text characteristics to perform feature extraction. In VETT algorithm, the iterative process was divided into two parts:vertical ensemble within the group and horizontal ensemble between groups. In-group ensemble is to construct an original classifier using the previous iterative models of the classifier, and the inter-group ensemble is to train three original classifiers through the traditional process to obtain the second-generation classifiers after this iteration, thereby improving the accuracy of the labels. Compared with Co-training, Tri-training, PU learning based on Area Under Curve (PU-AUC) and Vertical Ensemble Co-training (VECT) algorithms, VETT algorithm has the maximum value of F1 increased by 6.5, 5.08, 4.27 and 4.23 percentage points respectively. Experimental results show that the proposed VETT algorithm has better classification performance.

Focusing on the issue that traditional open-loop power control algorithm normally aims to increase the throughput and ignores the interference to other cells, to achieve a tradeoff between edge users and whole system performance, an Uplink Joint Power Control algorithm of LTE system (UJPC), was proposed. In the algorithm, single base station and three sectors were adopted as system model, which aimed to maximize proportional fair index of system throughput. Firstly, the corresponding mathematical optimization model was obtained according to two constraints of the minimum Signal-to-Interference plus Noise Ratio (SINR) and the maximum transmit power of users. Then continuous convex approximation method was used to solve optimization problem to get optimal transmission power of all users in each cell. The simulation results show that, compared with open-loop scheme, UJPC can greatly improve spectrum utilization of cell edge while ensuring average spectrum utilization of system and its best performance gain can reach 50%.

Considering the time delay and the energy consumption of terminal equipment caused by high-speed data transmission and calculation, a transmission scheme with equal power allocation in uplink was proposed. Firstly, based on collaborative properties of Augment Reality (AR) services, a system model for AR characteristics was established. Secondly, system frame structure was analyzed in detail, and the constraints to minimize total energy consumption of system were established. Finally, with the time delay and energy consumption constraints satisfied, a mathematical model of Mobile Edge Computing (MEC) resource optimization based on convex optimization was established to obtain an optimal communication and computing resource allocation scheme. Compared with user independent transmission scheme, the total energy consumption of the proposed scheme with a maximum time delay of 0.1 s and 0.15 s was both 14.6%. The simulation results show that under the same conditions, compared with the optimization scheme based on user independent transmission, the equal power MEC optimization scheme considering cooperative transmission between users can significantly reduce the total energy consumption of system.

The precision of user identification is low since the subjective weighting algorithms ignore the special meanings and effects of attributes in applications. To solve this problem, an Information Entropy based Multiple Social Networks User Identification Algorithm (IE-MSNUIA) was proposed. Firstly, the data types and physical meanings of different attributes were analyzed, then different similarity calculation methods were correspondingly adopted. Secondly, the weights of attributes were determined according to their information entropies, thus the potential information of each attribute could be fully exploited. Finally, all chosen attributes were integrated to determine whether the account pair was the matched one. Theoretical analysis and experimental results show that, compared with machine learning based algorithms and subjective weighting algorithms, the performance of the proposed algorithm is improved, on different datasets, the average precision of it is up to 97.2%, the average recall of it is up to 94.1%, and the average comprehensive evaluation metric of it is up to 95.6%. The proposed algorithm can accurately identify user accounts across multiple social networks.