Classification and regression algorithms based on k-nearest neighbors (kNN) are often ranked among the top-10 Machine learning algorithms, due to their performance, flexibility, interpretability, non-parametric nature...
详细信息
Classification and regression algorithms based on k-nearest neighbors (kNN) are often ranked among the top-10 Machine learning algorithms, due to their performance, flexibility, interpretability, non-parametric nature, and computational efficiency. Nevertheless, in existing kNN algorithms, the kNN radius, which plays a major role in the quality of kNN estimates, is independent of any weights associated with the training samples in a kNN-neighborhood. This omission, besides limiting the performance and flexibility of kNN, causes difficulties in correcting for covariate shift (e.g., selection bias) in the training data, taking advantage of unlabeled data, domain adaptation and transfer learning. We propose a new weighted kNN algorithm that, given training samples, each associated with two weights, called consensus and relevance (which may depend on the query on hand as well), and a request for an estimate of the posterior at a query, works as follows. First, it determines the kNN neighborhood as the training samples within the kth relevance-weighted order statistic of the distances of the training samples from the query. Second, it uses the training samples in this neighborhood to produce the desired estimate of the posterior (output label or value) via consensus-weighted aggregation as in existing kNN rules. Furthermore, we show that kNN algorithms are affected by covariate shift, and that the commonly used sample reweighing technique does not correct covariate shift in existing kNN algorithms. We then show how to mitigate covariate shift in kNN decision rules by using instead our proposed consensus-relevance kNN algorithm with relevance weights determined by the amount of covariate shift (e.g., the ratio of sample probability densities before and after the shift). Finally, we provide experimental results, using 197 real datasets, demonstrating that the proposed approach is slightly better (in terms of F-1 score) on average than competing benchmark approaches for mit
This study introduces an adaptive integral sliding mode disturbance observer (AISMDOB)-based robust bidirectional platoon control method, aiming to ensure mesh stability in vehicular systems. Most existing platoon con...
详细信息
This study introduces an adaptive integral sliding mode disturbance observer (AISMDOB)-based robust bidirectional platoon control method, aiming to ensure mesh stability in vehicular systems. Most existing platoon control studies only focus on error propagation stability in either the longitudinal or lateral direction, neglecting the uncertainties in kinematics and dynamics of vehicular systems. The study proposes new coupled spacing error dynamics derived from vehicle kinematics and extended look-ahead-based coupled spacing errors to ensure both the longitudinal and lateral error propagation stability (that is, mesh stability) and are subsequently utilized to develop the novel AISMDOB, which improves the existing integral sliding mode disturbance observers (ISMDOBs) by incorporating adaptive estimation of unknown disturbance bounds while preserving their advantages. The AISMDOB-based platoon control method is then proposed using both robust kinematic and dynamic controllers to effectively compensate for the kinematic disturbances and dynamic model uncertainties, thereby reducing chattering phenomenon and ensuring the asymptotic convergence of spacing and velocity errors. Additionally, the proposed method can prevent cutting-corner behaviors during cornering maneuvers by utilizing the coupled spacing error dynamics. Simulation and experimental results verify the effectiveness of the proposed method through comparison with ISMDOB-based, sliding mode control (SMC)-based, and previous extended look-ahead-based methods. IEEE
Use of multi-path network topologies has become a prominent technique to assert timeliness in terms of age of information (AoI) and to improve resilience to link disruptions in communication systems. However, establis...
详细信息
Network games provide a framework to study strategic decision making processes that are governed by structured interdependencies among agents. However, existing models do not account for environments in which agents s...
详细信息
This paper presents a lightweight and accurate convolution neural network (CNN) based on encoder in vision transformer structure, which uses multigroup convolution rather than multilayer perceptron and multiheaded sel...
详细信息
This paper first determines the generalized optical orthogonal code (GOOC) parameters to minimize the bit error probability in fiber-optic code division multiple access systems. The systems use on-off keying as the mo...
详细信息
True-time-delay (TTD) arrays can implement frequency-dependent rainbow beams and enable fast beam alignment in wideband millimeter-wave (mmWave) systems. In this paper, we consider 3D rainbow beam training with planar...
详细信息
Multiple access techniques are fundamental to the design of wireless communication systems,since many crucial components of such systems depend on the choice of the multiple access technique[1].For example,the use of ...
详细信息
Multiple access techniques are fundamental to the design of wireless communication systems,since many crucial components of such systems depend on the choice of the multiple access technique[1].For example,the use of orthogonal frequency-division multiple access (OFDMA) simplifies the physical layer design,
III-nitride nanowires have emerged as an important semiconductor device technology development platform, leveraging the unique physical properties of III-nitride semiconductors such as widely tunable bandgap energies,...
详细信息
Positioning clothing parts ($\mathcal {S} \mathbf{s}$) such as sleeves and collars has been in the realm of manual task that had to be done meticulously in order to prevent unnecessary tanglements during the simulatio...
详细信息
Positioning clothing parts ($\mathcal {S} \mathbf{s}$) such as sleeves and collars has been in the realm of manual task that had to be done meticulously in order to prevent unnecessary tanglements during the simulation. This paper proposes an optimization-based method to computerize the above $\mathcal {S}$-positioning task. For that, we embed each $\mathcal {S}$ to an abstracting cylinder $\mathcal {C}$ such that $\mathcal {S}$-positioning can be done by adjusting only 3$\sim$4 DOFs (e.g., translating/rotating $\mathcal {C}$ or adjusting its radius) instead of per-vertex-full-DOFs. Then, we formulate an objective function E by scoring undesirableness of $\mathcal {S}$'$\mathbf{s}$ position (e.g., $\mathcal {S}$ penetrating the body, $\mathcal {S}$ making cloth-to-cloth intersection). In organizing E into the loop of the Newton's method, the main challenge was to calculate the symbolic gradient and hessian, for which this paper makes several novel contributions. The resultant $\mathcal {S}$-positioning method works quite successfully;$\mathcal {S}$*$\mathbf{s}$ (the output of the S-positioning method ) are tanglement-free thus running the simulator to that configuration produces acceptable draping quickly;Experiments show that, in obtaining acceptable draping, the proposed method produces about ×9.7 speed up compared to when not using it. IEEE
暂无评论