Model selection (MS) and model averaging (MA) are two popular approaches when many candidate models exist. Theoretically, the estimation risk of an oracle MA is not larger than that MS because former more flexible, but a foundational issue this: Does offer substantial improvement over MS? Recently, seminal work by Peng Yang (2022) has answered this question under nested with linear orthonormal series expansion. In current paper, we further respond to regression models. A general framework,...

Smartphones are ubiquitous and becoming increasingly sophisticated, with ever-growing sensing powers. Recent years, more applications of activity recognition based on sensors developed for routine behavior monitoring helping the users form a healthy habit. In this field, finding an efficient method recognizing physical activities (e.g., sitting, walking, jogging, etc) becomes pivotal, core urgent issue. study, we construct Convolutional Neural Network (CNN) to identify human using data...

In financial econometrics, the Sharpe ratio function serves as a gold standard to measure return-to-risk for comparing different assets or trading strategies. recent literature, several methods have been developed directly indirectly estimate function, yet none of them apply scenario where covariates are measured with error. To handle this problem, we propose new method by incorporating local polynomial smoothing and SIMEX simultaneously negative log-volatility in presence measurement The...

Since the implementation of Basel III Accord, expected shortfall (ES) has gained increasing attention from regulators as a complement to value-at-risk (VaR). The problem elicitability for ES makes jointly modeling VaR and popular method study ES. In this article, we develop model averaging joint regression models that selects two weight vectors by minimizing jackknife criterion. We show large sample properties estimators under potential misspecification with dimension parameters asymptotic...

Sequential recommendation is an advanced technique that utilizes the sequence of user behaviors to generate personalized suggestions by modeling temporal dependencies and patterns in preferences. However, it requires a server centrally collect users' data, which poses threat data privacy different users. In recent years, federated learning has emerged as distributed architecture allows participants train global model while keeping their private locally. This survey pioneers Federated...

Many unsupervised visual anomaly detection methods train an auto-encoder to reconstruct normal samples and then leverage the reconstruction error map detect localize anomalies. However, due powerful modeling generalization ability of neural networks, some anomalies can also be well reconstructed, resulting in unsatisfactory localization accuracy. In this paper, a small coarsely-labeled dataset is first collected. Then, coarse-knowledge-aware adversarial learning method developed align...

The proliferation of powerful mobile devices, along with the increasing demand location information, has driven researches to develop positioning applications higher accuracy. In this paper, we build an Android application which achieved aim seamless between inside and outside. system contains four parts: outdoor positioning, uses GPS Baidu Map, indoor adopts Wi-Fi fingerprint contextual detection, including detection floor takes advantage multiple sensors integrated on Android, functional...

Satellite networks are capable of contenting a variety data transmission needs users in geographically diverse locations throughout the world. Multi‐layered satellite (MLSNs) can construct efficient communications due to their extensive coverage and high network capacity. However, throughput degradation severe end‐to‐end delay could occur MLSNs because traffic congestion. To resolve these problems, authors first propose novel MEO/LEO architecture that effective inter‐satellite links. Then...

Nowadays smartphones are equipped with various sensors and powerful processing modules, accessible to flexible communication networks, thus enabling complex applications such as context awareness, activity recognition, health care monitoring so forth. These typically require contextual information optimize the effectiveness, e.g. indoor/outdoor identification. This paper develops an detection method based on a generic smartphone platform, utilizing extracted from internal clock, GPS module...

With the increasing development of technologies and reducing costs materials, satellite networks have been widely investigated both on demand supply for many years. Multi-Path routing has applied as an effective method to support end-to-end reliability. But there exist inter-satellite links loss long delay in current multi-path due much frequency route request flooding. In this paper, improved algorithm (IMP) hierarchical Medium Earth Orbit (MEO) Low (LEO) satellites is proposed evaluated....

Channel interference and node mobility cause significant performance degradation to wireless networks. In multichannel multi-flow mobile cognitive ad hoc networks, it becomes even worse due both unexpected primary user activities potential among multiple flows. this paper, we propose a Joint Routing Assignment (JRCA) approach based on delay prediction. Firstly, formulates the JRCA problem with objective of minimization. Next, prediction model is proposed channel collision probability. Then,...

The production of autonomous vehicles needs to accumulate a large amount mileage experience while the researching, development and testing methods traditional vehicle terminals cannot meet requirements. industry generally adopts new Cloud Platform research develop iterative mode, which collects massive road environment data through test fleet, conducts model algorithm simulation verification, so as product safety standards. Based on collaboration capability Cloud, Pipe Terminal, this paper...

Software Defined Networking (SDN) provides a flexible way to control heterogeneous network resources, which can effectively manage the space and terrestrial resources. The controller placement problem is important for SDN-based network, it an entirely new challenge in space-terrestrial integrated networks (STINs). In this paper, we firstly propose STIN architecture, where integrates resources layers. Then, formulate dynamic proposed architecture as mixed non-integer linear programming prove...

This paper introduces a framework of contextual thinking in mobile devices. It is based on real-time sensing local time, significant locations, location-dwelling states, and user states to infer activities. A activity well-defined be inferred, for example, waiting bus, having meeting, working office, taking break coffee shop et al. location defined as geofence, which can node associated with circle, or polygon. state enter into location, the duration, exit from location. combination mobility...

Data transmission through satellite networks provides ubiquitous communication all over the world. But time-variant topologies lead to lossy inter-satellite links and long end-to-end delay. To address these problems, in this paper we propose a novel network management method routing algorithm. The former specializes satellites different layers utilize resources efficiently, latter is an optimized layered algorithm based on TORA Dijkstra's algorithms. proposed customized for hybrid GEO/LEO...

Path planning and control logic are the focus of research in Driverless Formula Racing key to improving lap times. Vehicle kinematics dynamics models designed based on design parameters driverless formula racing vehicle competition Student Autonomous China (FSAC). The path is accomplished track data collected by sensors through software OpenLap using method Model Predictive Control (MPC). established model strategy simulated Driving Scenario Designer Matlab/Simulink gazebo Robot Operating...

Most existing routing algorithms assume wireless nodes use maximal transmission power or set up the at beginning of network configuration. These static approaches potentially introduce signal interference that can be mitigated through control. In this paper, we propose a Joint Channel assignment, stable Routing and adaptive Power control (JCRP) approach dynamically controls to avoid channel for improving utility. Our JCRP allows node its certain value which it has longest conflict-free time....

The benefits of network layer cooperation cognitive radio networks have been gradually recognized in recent years. In this paper we consider the network, whereby primary users select some secondary to relay packets, return for more favourable spectrum access rules users. Under scheme, investigate how enlarge throughput whole where a QoS(Quality Service)-guaranteed adaptive scheme is developed. Our can guarantee QoS demand and update its frequency division dynamically according status nodes....

The hierarchical SDN network can effectively integrate the heterogeneous resources of space-terrestrial integrated networks through its reconfigurable management capabilities. In a network, how to achieve controllers is very important issue. this paper, we first construct SDN-based STINs architecture resources. Then, extend standard southbound interface OpenFlow protocol manage Finally, design Multipath transmission algorithm (MTA) based on information provided by extended link load...

The main research content of this paper is vehicle longitudinal and lateral control complex scheduling problem platoon. For the navigation module single control, Gazebo used to simulate real environment, 2D lidar, binocular camera tag codes are carry out localization navigation. multi-vehicle part, mainly pays attention switching platoon states. One repeated switch convoy from formation disbandment, two strategies adopted, namely, disbandment state strategy strategy. In order achieve...