A great deal of research has focused on using convolutional neural network for optical character recognition. However we encountered two typical problem in this field when applied convolutional neural network to handw...
详细信息
ISBN:
(纸本)9781538632215
A great deal of research has focused on using convolutional neural network for optical character recognition. However we encountered two typical problem in this field when applied convolutional neural network to handwritten Yi character recognition. First, since convolutional neural network is a kind of supervised deep learning model, the manual training data labeling is a very time consuming and labor intensive work. Second, because the theory is not well studied, the structure design and parameter adjustment of convolutional neural network depend heavily on experience, and our recognition accuracy was not satisfactory at the beginning. To address these two problems, in this paper, for one thing, we use entropy theory improved a density-based clustering algorithm, which is proved very effective in data labeling. For another, as to the problem of structure design and parameter adjustment, we compared performance of models with different scales and different parameters, and gave some experience about this problem. Finally we achieved 99.65% accuracy on the test set. We hope that this paper will inspire more researches on convolutional neural network applied to dataset-lacked optical character recognition problems.
Cognitive radio is a key technology for promoting spectrum efficiency by exploiting the existence of spectrum holes under the current static spectrum allocation policy. However, using spectrum bands in an opportunisti...
详细信息
ISBN:
(纸本)9783319092652;9783319092645
Cognitive radio is a key technology for promoting spectrum efficiency by exploiting the existence of spectrum holes under the current static spectrum allocation policy. However, using spectrum bands in an opportunistic way has brought some significant challenges such as connectivity and energy consumption in dynamic environment. In this paper, we propose an adaptive SDEC (spectrum-aware degree-ranking-based energy-efficient clustering algorithm) applying to multi-hop wireless CRN, which incorporates the spectrum quality and the number of neighbor nodes into consideration. Results of simulations illustrate that the algorithm can well fit into CR network and achieve significant improvement both on the load balancing and the average power consumption. Moreover, SDEC has preferable stability and validity due to its low complexity and quick convergence under dynamic spectrum change.
This paper takes the logistics distribution record of Yifeng Weiye Group for the past two years as the basic research unit. By exploring the relationship between data fields, we use the idea of adaptive clustering alg...
详细信息
ISBN:
(纸本)9781450360920
This paper takes the logistics distribution record of Yifeng Weiye Group for the past two years as the basic research unit. By exploring the relationship between data fields, we use the idea of adaptive clustering algorithm and spatial clustering analysis to process the attribute data of transportation capacity[8]. Basing on the obtained clustering results, we use Python and PHP technology to optimize the distribution area, and finally design an effective visual expression method to obtain the traffic situation knowledge. We can provide relevant analysis and technical support for enterprises to improve the efficiency of distribution logistics and optimize the structure of the industrial chain.
In ad-hoc networks, MSWCA is a typical algorithm in clustering algorithms with consideration on motion-correlativity. Aiming at MSWCA's problem that "it only considers on intra-cluster stability, and neglects...
详细信息
ISBN:
(纸本)9783038351153
In ad-hoc networks, MSWCA is a typical algorithm in clustering algorithms with consideration on motion-correlativity. Aiming at MSWCA's problem that "it only considers on intra-cluster stability, and neglects the inter-cluster stability", a new clustering algorithm (NCA) was proposed. Firstly, NCA clustering algorithm and its cluster maintenance scheme were designed. Secondly, the theoretical quantitative analyses on average variation frequency of clusters and clustering overheads were conducted. The results show that NCA can improve cluster stability and reduce clustering overheads.
In clustered heterogeneous networks, some nodes are set as cluster heads, responsible to integrate the information from the intra-cluster members and send it to the sink node. Therefore cluster heads dissipate much mo...
详细信息
ISBN:
(纸本)9781479925384
In clustered heterogeneous networks, some nodes are set as cluster heads, responsible to integrate the information from the intra-cluster members and send it to the sink node. Therefore cluster heads dissipate much more energy than other sensor nodes. And the stability of networks can be affected when the first node dies. We assume that nodes are assumed be equipped with different energy randomly, a clustering algorithm based on residual energy and multi-management for cluster heads is proposed. By adding a parameter into the probability of cluster head election, the proposed algorithm makes nodes with more initial energy and residual energy become cluster heads with large probability. It balances the energy consumption of network and prolongs the survival time of network. At the same time, in order to ensure the quality of network transmission, the strategy of multi-hop management is introduced in this algorithm. We show by simulation that the algorithm has a longer lifetime and more stable capacity of data transmission than LEACH, DEEC and SEP in the multi-level energy heterogeneous network.
A hybrid clustering algorithm based on the artificial immune theory is presented in this paper. The method is inspired by the clone selection and memory principle. The problem of local optimal can be avoided by introd...
详细信息
ISBN:
(纸本)9783642247279
A hybrid clustering algorithm based on the artificial immune theory is presented in this paper. The method is inspired by the clone selection and memory principle. The problem of local optimal can be avoided by introducing the differentiation of memory antibody and inhibition mechanism. In addition, the K-means algorithm is used as a search operator in order to improve the convergence speed. The proposed algorithm can obtain the better data convergence compared with the K-means algorithm based clustering approach and artificial immune based approach. Simulate experimental results indicate the hybrid algorithm has a faster convergence speed and the obtained clustering centers can get strong stability.
A low-complexity clustering algorithm can achieve modulation recognition and resistant nonlinearity distortion of RoF System is proposed. This technique can classify different modulations accurately for any order and ...
详细信息
ISBN:
(纸本)9781538691458
A low-complexity clustering algorithm can achieve modulation recognition and resistant nonlinearity distortion of RoF System is proposed. This technique can classify different modulations accurately for any order and can improve the BER performance.
With the energy constrained nature of wireless sensors, it is a substantial design issue to make efficient use of battery power in order to increase their lifetime. Focuses on reducing energy consumption of wireless s...
详细信息
ISBN:
(纸本)9780769535227
With the energy constrained nature of wireless sensors, it is a substantial design issue to make efficient use of battery power in order to increase their lifetime. Focuses on reducing energy consumption of wireless sensor network, this paper proposed CABCF-DCS (clustering algorithm based on communication facility with deterministic cluster-size) algorithm. By changing the cluster-size of the cluster based on CABCF (clustering algorithm based on communication facility) algorithm, the new algorithm achieves the purpose of saving energy ultimately. Simulation results validate the energy efficiency of the new algorithm.
On mountain highways, excessive speeding appears to be the major factor that causes fatal traffic accidents. In order to find out the most significant attributes that affects the vehicle speed so that countermeasures ...
详细信息
ISBN:
(纸本)9781424469284
On mountain highways, excessive speeding appears to be the major factor that causes fatal traffic accidents. In order to find out the most significant attributes that affects the vehicle speed so that countermeasures could be applied. Therefore, based on the actual traffic data collected from black spots in four provinces of southwestern China;this paper investigated the significant roadway attributes that influence the speed of small cars on mountain highways through Two step clustering algorithm, which is capable of clustering both categorical and continuous attributes. Through experiments, it shows that "material of roadway" and "presence of protective facilities at roadside" are two major attributes that affect the vehicle speed;furthermore, "material of roadway" contributes relatively more on affecting vehicle speed.
Traditional fingerprint orientation clustering algorithms often use k means clustering algorithm, but as a result of fingerprint and objective factors of volatile characteristics over time, k-means cannot adapt to cha...
详细信息
ISBN:
(纸本)9783319745213;9783319745206
Traditional fingerprint orientation clustering algorithms often use k means clustering algorithm, but as a result of fingerprint and objective factors of volatile characteristics over time, k-means cannot adapt to change at any time in fingerprint, and cannot be generated adaptive clustering cluster number, cause the matching accuracy is not high. This paper adopts a based on support vector machine (SVM) and DBSCAN clustering algorithm, can generate continuously adapt to changing the optimal hyperplane fingerprint model, solved the fingerprint fluctuating lead to the problem of matching result is bad, and can be automatically generated in the process of matching classification number of clusters, based on statistical density characteristics of DBSCAN selection matching probability model, to improve the positioning of the matching accuracy, reduced the amount of time matching positioning, positioning accuracy can be up to 2.04 m in the range of 57%, relative k-means 6.1 m increased by 52.3%, improve the positioning accuracy.
暂无评论