Hypertension is an important hazard to health. Blood pressure prediction is of great importance to avoid grave consequences caused by sudden increase of blood pressure. Based on traditional Long Short-Term Memory (LSTM) network, a multi-factor cue LSTM model for both short-term prediction (predicting blood pressure for the next day) and long-term prediction (predicting blood pressure for the next several days) was proposed to provide early warning of undesirable change of blood pressure. Multi-factor cues used in blood pressure prediction model included time series data cues (e.g. heart rate) and contextual information cues (e.g. age, BMI (Body Mass Index), gender, temperature).The change characteristics of time series data and data features of other associated attributes were extracted in the blood pressure prediction. Environment factor was firstly considered in blood pressure prediction and multi-task learning method was used to help the model to capture the relation between data and improve the generalization ability of the model. The experimental results show that compared with traditional LSTM model and the LSTM with Contextual Layer (LSTM-CL) model, the proposed model decreases prediction error and prediction bias by 2.5%, 3.8% and 1.9%, 3.2% respectively for diastolic blood pressure, and reduces prediction error and prediction bias by 0.2%, 0.1% and 0.6%, 0.3% respectively for systolic blood pressure.

The traditional literatures mainly consider the cooperative relationship while neglecting the competitive relationship between manufacturing and logistics enterprises during interactive development. An improved model, namely Logistic-Volterra model, was proposed based on the traditional Logistic model, which considered the contribution coefficients and competition coefficients at the same time. Firstly, the Logistic-Volterra model was built and the stability solution was sovled, then the mathematical conditions for achieving stability and the interpretation of reality were discussed. Secondly, the affecting factors on the interactive development of manufacturing and logistics enterprises were discovered by using Matlab numerical simulation, and the differences between the improved model and traditional model were also discussed. Finally, the manufacturing enterprise A and logistics enterprise B were taken as an example to analyze the competitive behavior in the process of cooperation; furthermore, the impact of coopetition behavior on the interests was also analyzed. The theoretical analysis and simulation results show that the stability of the system is highly affected by contribution coefficient, competition coefficient and environmental capability, the result is more reasonable when considering the competition relationship in the model. It means that manufacturing and logistics enterprises should fully address the effects of competition on both sides.

To solve the problem of complex context and data sparsity, a new algorithm for the tensor decomposition based on context similarity of mobile user was proposed, namely UCS-TF (User-Context-Service Tensor Factorization recommendation). Multi-dimensional context similarity model was established with combining the user context similarity and confidence of similarity. Then, K-neighbor information of the target user was applied to the three-dimensional tensor decomposition, composed by user, context and mobile-service. Therefore, the predicted value of the target user was obtained, and the mobile recommendation was generated. Compared with cosine similarity method, Pearson correlation coefficient method and the improved Cosine1 model, the Mean Absolute Error (MAE) of the proposed UCS-TF algorithm was reduced by 11.1%, 10.1% and 3.2% respectively; and the P@N index of it was also significantly improved, which is better than that of the above methods. In addition, compared with Cosine1 algorithm, CARS2 algorithm and TF algorithm, UCS-TF algorithm had the smallest prediction error on 5%, 20%, 50% and 80% of data density. The experimental results indicate that the proposed UCS-TF algorithm has better performance, and the user context similarity combining with the tensor decomposition model can effectively alleviate the impact of score sparsity.

Current research on statistical publication of differential privacy data stream only considers one-dimensional data stream. However, many applications require privacy protection publishing two-dimensional data stream, which makes traditional models and methods unusable. To solve the issue, firstly, a differential privacy statistical publication algorithm for fixed-length two-dimensional data stream, call PTDSS, was proposed. The tuple frequency of the two-dimensional data stream under certain condition was calculated by a one-time linear scan to the data stream with low-cost space. Basing on the result of sensitivity analysis, a certain amount of noise was added into the statistical results so as to meet the differential privacy requirement. After that, a differential privacy continuous statistical publication algorithm for any length two-dimensional data stream using sliding window model, called PTDSS-SW, was presented. The theoretical analysis and experimental results show that the proposed algorithms can safely preserve the privacy in the statistical publication of two-dimensional data stream and ensure the relative error of the released data in the range of 10% to 95%.

To solve the problem of optimization learning models in Belief Rule Base (BRB), a new parameter training approach based on the Particle Swarm Optimization (PSO) algorithm was proposed, which is one of the swarm intelligence algorithms. The optimization learning model was converted to nonlinear optimization problem with constraints. During the optimization process, all particles were limited in the search space and the particles with no speed were given velocity in order to maintain the diversity of the population of particles and achieve parameter training. In the practical pipeline leak detection problem, the Mean Absolute Error (MAE) of the trained system was 0.166478. The experimental results show the proposed method has good accuracy and it can be used for parameter training.

Under the condition of being confronted with highly concurrent requests, the existing Web services would bring about the increase of the response time, even the problem that server goes down. To solve this problem, a kind of distributed self-elasticity architecture for the Web system named ECAP (self-Elasticity Cloud Application Platform) was proposed based on cloud computing. The architecture built on the Infrastructure as a Service (IaaS) platform of OpenStack. It combined Platform as a Service (PaaS) platform of Cloudify to realize the ECAP. In addition, it realized the fuzzy analytic hierarchy scheduling method by building the fuzzy matrix in the scale values of virtual machine resource template. At last, the test applications were uploaded in the cloud platform, and the test analysis was given by using the tool of pressure test. The experimental result shows that ECAP performs better in the average response time and the load performance than that of the common application server.

With the growth of cloud technology, distributed application platform develops towards elasticity resources and dynamic migration environment. The refinement of distributed application access control policies was associated with resources and environment, which also needs to improve performance to adapt to the dynamics. Although present access control space policies conflict analysis methods could be used in the conflict analysis of distributed application access control policies refinement. The granularity of its calculating unit is too fine to make batter performance. In this article, the authors designed a conflict analysis algorithm used in distributed application access control policies refinement, the conflict analysis algorithm was based on recursive calculation the intersection of sets and the calculation unit of the algorithm was permission assignment unit which improved computing granularity. The experimental results and analysis show that the proposed algorithm has better performance, and fits the needs of improving computing performance of cloud platform access control policies refinement.

In system call trace of a process, there are two kinds of invariability, program behavior invariability and user behavior invariability, of which the former can be further subdivided into temporal order invariability and frequency invariability. The existing researches on system call based intrusion detection techniques focus on program behavior invariability only, ignoring user behavior invariability. Based on frequency invariability embedded in process traces, the existence and description of user behavior invariability were studied, on which the least entropy length was proposed to measure the invariability. The experiment on Sendmail datasets shows that, least entropy length excellently describes user behavior invariability and significantly improves the performance of system call anomaly detection with the help of program behavior invariability.

Based on Brightness preserving Bi-Histogram Equalization (BBHE), an improved algorithm of gray image enhancement was proposed. An appropriate threshold, which was selected based on the entropy of the output image and the difference between the mean brightness of input and output images, was selected to cut the image, then BBHE and gray lever homogenization were performed respectively. This method makes the brightness mean error as small as possible and the entropy of output image as large as possible. Meanwhile, it can prevent over-enhancement. The experimental results prove that the new method has better performance on image enhancement.

In order to align partly overlapped data clouds measured from different view points, a multi-scale feature points detection algorithm was proposed. A few feature points can be extracted from large number of original data quickly. This algorithm consists of three steps: discrete curvature computing, bilateral filtering and feature points detecting. The number of feature points can be controlled by scale parameter approximately. For each feature point, the authors proposed local shape spectral descriptor to identify its local shape characteristic. Firstly, an affinity matrix was constructed using distance and curvature information of points in neighborhood of a feature point, and then a few of eigenvalues of affinity matrix were used to form a shape descriptor, with which the correspondence between different data sets can be computed easily. Some examples prove that the method is robust and efficient for aligning large number of data with noise.