Papers by Keyword: Clustering

Abstract: This paper combines the data mining technology and the rich sports consumption data resources of city household survey. By using the K-Means fast cluster method, sports consumer market models were constructed based on the different variables. Research shows, choosing sports consumption content as variables to establish clustering model is better than choosing the demographic , sports consumption content ,consumer psychology and way of life as variables to establish clustering model. According to the results of clustering, the city residents are divided into four kinds of consumer groups in accordance with the different features of sports consumption.

Abstract: Presented paper describes vision-based algorithm for 2D Rubik’s cube state detection suitable for cases in which long-term fixed camera-cube position is not possible. The main focus was to provide a robust algorithm for position and color detection in order to overcome problems observed in previous version. First part of paper describes Hough transform and advanced clustering functions that were used for cube position detection. Described algorithm provides robustness to strong occlusions and various lighting conditions. Second part of paper describes color detection algorithm and problems of prior classification in various color spaces.

Abstract: Mountainous environment makes the wireless sensor network (WSN) data collection applications remain a challenging domain, as the detection region may present a three-dimensional structure and the radio propagation characteristics are still looking forward to further research. To better adapt to the ecological data acquisition needs in mountainous orchard, A WSN clustering data acquisition system is designed and implemented. It uses the received signal strength indicator (RSSI) to evaluate radio propagation performance and characterize the communication quality of the link. In the selection of the cluster heads and the next routing hops, this system takes RSSI, node’s residual energy and other influencing factors into account and use the multiple-attribute comprehensive evaluation model to weigh them comprehensively. Simulation results indicate that such a design can give objective and reasonable evaluations and judgments of the candidate nodes. Analyses verify the effectiveness and reasonability of the proposed model.

Abstract: The k anonymity was one of the first algorithms applied for privacy protection in location-based service(LBS).The k anonymity exhibits its disadvantages gradually, such as being easily attacked by continuous queries attacking algorithm, the larger k value for higher security level lead to more pointless cost of bandwidth and load of LBS server. This article analyzes the causes of the problems, and proposes a new idea based on clustering algorithm to improve the k anonymity algorithm.

Abstract: In this paper, a new approach of genetic algorithm called knowledge-based Genetic Algorithm (KBGA-Clustering) is proposed for multidimensional data clustering . Basically, this method adopts knowledge of what called as appropriate cluster centre for a fixed number of k-cluster. The chromosome which has inappropriate genes will be penalised with maximum value to prohibit it in the next generation. The experimental result is also provided for KBGA-Clustering and Genetic Algorithm-Clustering (GA-Clustering) to present the performance. Based on the observation, KBGA-Clustering presents better performance and more optimum solution compared to conventional GA-Clustering.

Abstract: A topic discovery approach of the image has been proposed. First, the training images are segmented into some blocks. After clustering blocks, we obtained class set generated by cluster centers, and extracted all nouns from text annotation of each training image to obtain a keyword set. Secondly, the un-label testing image is also segmented into some blocks as same as training images, we calculated the correlation between the block and keyword, and the keyword set for each block may be obtained. Finally, the number of the same keyword appearing in the each block is calculated, we let the keywords with maximum to appear times be as the topics of the image. The experimental results confirm that proposed approach for the image is effectiveness and has good performance.

Abstract: MCL is a graph clustering algorithm. With the characteristics of the MCL computational process, MCL is prone to producing small clustering and separating edge nodes from the group. A hybrid clustering based on MCL combined with KNN algorithm is proposed. Hybrid algorithm improves the quality of clustering by reclassification of elements in small clustering by using KNN classification characteristics and Clustering tables required by MCL clustering. Experiment proves the improved algorithm can enhance the quality of clustering.

Abstract: Clustering algorithm can effectively solve the imbalance of energy consumption of different nodes. Based on the analysis of traditional LEACH protocol, we propose an improved sensor network clustering routing protocol, which reduced the cost of control information in formation process of clusters by adopting a new kind of competition parameters of cluster head, so as to solve the problem of energy heterogeneous of network node. The simulation results show that the protocol can effectively save the energy consumption of the nodes, and prolong the network life time.

Abstract: A topic discovery approach of the image has been proposed. First, the training images are segmented into some blocks. After clustering blocks, we obtained class set generated by cluster centers, and extracted all nouns from text annotation of each training image to obtain a keyword set. Secondly, the un-label testing image is also segmented into some blocks as same as training images, we calculated the correlation between the block and keyword, and the keyword set for each block may be obtained. Finally, the number of the same keyword appearing in the each block is calculated, we let the keywords with maximum to appear times be as the topics of the image. The experimental results confirm that proposed approach for the image is effectiveness and has good performance.

Abstract: This paper proposes an energy-efficient clustering routing algorithm based on the node degree, relative distance between nodes and residual energy. Selecting the cluster head fully consider the node degree and their relative distance. By this way, the coverage performance of cluster head is better and the average distance between member nodes and cluster head is short, so the cost of communication in the cluster is decreased. At the same time, the nodes with low energy have less probability to become the cluster head. Therefore, it can improve the quality of cluster; prolong the life time of the network.