Knowledge bases (KBs) are often greatly incomplete, necessitating a demand for KB completion. The path ranking algorithm (PRA) is one of the most promising approaches to this task. Previous work on PRA usually follows single-task learning paradigm, building prediction model each relation independently with its own training data. It ignores meaningful associations among certain relations, and might not get enough data less frequent relations. This paper proposes novel multi-task framework...

The uncertainty rock mechanical parameters (i.e., deformation and strength parameters) is an important factor in the safety estimation support design of underground engineering. Ignoring this could allow potential risks to structure. To address challenge, paper develops verifies a Bayesian approach for rock's by integrating limited site data prior knowledge, integrated knowledge then transformed into large number equivalent samples parameters. experimental marble from triaxial compression...

Real-time prediction of excavation-induced displacement retaining pile during the deep excavation process is crucial for construction safety. This paper proposes a modified back analysis method with multi-objective optimization procedure, which enables real-time horizontal construction. As opposed to traditional stage-by-stage analysis, time series monitoring data till current stage are utilized form function. Then, particle swarm (MOPSO) algorithm applied parameter identification. The...

In this paper, a joint non-orthogonal multiple access and time division (NOMA-TDMA) scheme is proposed in Industrial Internet of Things (IIoT), which allowed sensors to transmit the same time-frequency resource block using NOMA. The user scheduling, slot allocation, power control are jointly optimized order maximize system α -fair utility under constraint minimum rate constraint. optimization problem nonconvex because fractional objective function constraints. To deal with original problem,...

To improve the ability of sine cosine algorithm (SCA) in exploitation process, an improved symmetric SCA with adaptive probability selection (SSCA-APS), is proposed. The search process this divided into early and late stages. In stage, operators traditional continue to be used. three improvements were applied. Firstly, are strategy adopted integrate original for dynamically adjusting step size range. Furthermore, prevent population from falling local optimization, Gaussian perturbation used...

Aiming at the problem of educational news classification, this paper proposes a method classification based on emotional dictionary. This takes as application background. First, summarizes common sentiment dictionary and education-related words manually arranged. Then, uses Laplace smoothing SO-PMI algorithm to expand news. Finally, whole is divided into several clauses, from which key in education required by combining dictionary, calculate weights each clause according corresponding rules...

The multiobjective evolutionary algorithm based on decomposition (MOEA/D) has received attention from researchers in recent years. This paper presents a new and the cloud model called Cloud Particle Differential Evolution (MOEA/D-CPDE). In proposed method, best solution found so far acts as seed each generation evolves two individuals by generator. A individual is produced updating current with position vector difference of these individuals. performance carried 16 well-known multi-objective...

Convolutional Neural Networks (CNNs) have demonstrated their superiority in image classification, and evolutionary computation (EC) methods recently been surging to automatically design the architectures of CNNs save tedious work manually designing CNNs. In this paper, a new hybrid differential evolution (DE) algorithm with newly added crossover operator is proposed evolve any lengths, which named DECNN. There are three ideas DECNN method. Firstly, an existing effective encoding scheme...

Fingerprint location algorithm is often used in process of indoor location. But the complex environment will cause fluctuations signal strength and this fluctuation lead to positioning errors. In order correct error we use Gaussian filtering preprocess data adopt based on voting region decision policy maximum AP selection policy. SSCM-WKNN that integrates these policies WKNN K-value optimized are implemented a simulation environment, experiments show it effective reduce error.

We propose a new optimization algorithm inspired by the formation and change of cloud in nature, referred to as Cloud Particles Differential Evolution (CPDE) algorithm. The is assumed have three states proposed Gaseous state represents global exploration. Liquid intermediate process from exploration local exploitation. Solid best solution found so far acts nucleus. In gaseous state, nucleus leads population explore condensation operation. liquid particles carry out macrolocal exploitation...

Data interpolation is crucial for 3D multiple predictions, both surface related elimination (SRME) and wave field extrapolation (WFE); particularly wide azimuth survey (WAZ). Combining the data with SRME WFE helps to improve prediction quality. regularization also RTM imaging without any additional cost.

In a vehicular ad hoc network (VANET), roadside units (RSUs) are installed at and intersections to process vehicle‐to‐infrastructure communication, collect analyse intelligent vehicle traffic data, send information vehicles, achieve early warning of safe driving vehicles. Owning the high cost implementing maintaining RSUs, it is vital importance determine where how many RSUs deploy. Optimal RSU deployment requires both small number maximum coverage running process, which constitutes...