• Journal of Electrical Engineering and Technology
• /
• v.13 no.5
• /
• pp.1841-1851
• /
• 2018

For the safe and stable operation of the power system, accurate wind power prediction is of great significance. A wind power prediction method based on empirical mode decomposition and improved extreme learning machine is proposed in this paper. Firstly, wind power time series is decomposed into several components with different frequency by empirical mode decomposition, which can reduce the non-stationary of time series. The components after decomposing remove the long correlation and promote the different local characteristics of original wind power time series. Secondly, an improved extreme learning machine prediction model is introduced to overcome the sample data updating disadvantages of standard extreme learning machine. Different improved extreme learning machine prediction model of each component is established. Finally, the prediction value of each component is superimposed to obtain the final result. Compared with other prediction models, the simulation results demonstrate that the proposed prediction method has better prediction accuracy for wind power.

• Journal of Korean Society for Atmospheric Environment
• /
• v.9 no.3
• /
• pp.207-215
• /
• 1993

In order to statistically predict $O_3$ levels in Seoul, the study used the TMS (telemeted air monitoring system) data from the Department of Environment, which have monitored at 20 sites in 1989 and 1990. Each data in each site was characterized by 6 major criteria pollutants ($SO_2, TSP, CO, NO_2, THC, and O_3$) and 2 meteorological parameters, such as wind speed and wind direction. To select proper variables and to determine each pollutant's behavior, univariate statistical analyses were extensively studied in the beginning, and then various applied statistical techniques like cluster analysis, regression analysis, and expert system have been intensively examined. For the initial study of high level $O_3$ prediction, the raw data set in each site was separated into 2 group based on 60 ppb $O_3$ level. A hierarchical cluster analysis was applied to classify the group based on 60 ppb $O_3$ into small calsses. Each class in each site has its own pattern. Next, multiple regression for each class was repeatedly applied to determine an $O_3$ prediction submodel and to determine outliers in each class based on a certain level of standardized redisual. Thus, a prediction submodel for each homogeneous class could be obtained. The study was extended to model $O_3$ prediction for both on-time basis and 1-hr after basis. Finally, an expect system was used to build a unified classification rule based on examples of the homogenous classes for all of sites. Thus, a concept of high level $O_3$ prediction model was developed for one of $O_3$ alert systems.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.10
• /
• pp.1712-1732
• /
• 2011

Cloud provides dynamically scalable virtualized computing resources as a service over the Internet. To achieve higher resource utilization over virtualization technology, an optimized strategy that deploys virtual machines on physical machines is needed. That is, the total number of active physical host nodes should be dynamically changed to correspond to their resource usage rate, thereby maintaining optimum utilization of physical machines. In this paper, we propose a pattern-based prediction model for resource provisioning which facilitates best possible resource preparation by analyzing the resource utilization and deriving resource usage patterns. The focus of our work is on predicting future resource requests by optimized dynamic resource management strategy that is applied to a virtualized data center in a Cloud computing environment. To this end, we build a prediction model that is based on user request patterns and make a prediction of system behavior for the near future. As a result, this model can save time for predicting the needed resource amount and reduce the possibility of resource overuse. In addition, we studied the performance of our proposed model comparing with conventional resource provisioning models under various Cloud execution conditions. The experimental results showed that our pattern-based prediction model gives significant benefits over conventional models.

• ETRI Journal
• /
• v.38 no.5
• /
• pp.807-817
• /
• 2016

A novel fast algorithm is suggested for a coding unit (CU) mode decision using pseudo rate-distortion based on a separated encoding structure in High Efficiency Video Coding (HEVC). A conventional HEVC encoder requires a large computational time for a CU mode prediction because prediction and transformation procedures are applied to obtain a rate-distortion cost. Hence, for the practical application of HEVC encoding, it is necessary to significantly reduce the computational time of CU mode prediction. As described in this paper, under the proposed separated encoder structure, it is possible to decide the CU prediction mode without a full processing of the prediction and transformation to obtain a rate-distortion cost based on a suitable condition. Furthermore, to construct a suitable condition to improve the encoding speed, we employ a pseudo rate-distortion estimation based on a Hadamard transformation and a simple quantization. The experimental results show that the proposed method achieves a 38.68% reduction in the total encoding time with a similar coding performance to that of the HEVC reference model.

• International conference on construction engineering and project management
• /
• 2009.05a
• /
• pp.203-211
• /
• 2009

Korean public owners who order public multi-family housing construction projects have yet to gain access to a model for predicting construction cost. For this reason, their construction cost prediction is mainly dependent upon historic data and experience. In this paper, a cost-prediction model based on Case-Based Reasoning (CBR) in the design phase of public multi-family housing construction projects was developed. The developed model can determine the total construction cost by estimating the different Building, Civil, Mechanical, Electronic and Telecommunication, and Landscaping work costs. Model validation showed an accuracy of 97.56%, confirming the model's excellent viability. The developed model can thus be used to predict the construction cost to be shouldered by public owners before the design is completed. Moreover, any change orders during the design phase can be immediately applied to the model, and various construction costs by design alternative can be verified using this model. Therefore, it is expected that public owners can exercise effective design management by using the developed cost prediction model. The use of such an effective cost prediction model can enable the owners to accurately determine in advance the construction cost and prevent increase or decrease in cost arising from the design changes in the design phase, such as change order. The model can also prevent the untoward increase in the duration of the design phase as it can effectively control unnecessary change orders.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.4
• /
• pp.463-468
• /
• 2011

Context prediction methods have been developed in two ways - one is a prediction for discrete context and the other is for continuous context. As most of the prediction methods have been used with prediction algorithms in specific domains suitable to the environment and characteristics of contexts, it is difficult to conduct a prediction for a user's context which is based on various environments and characteristics. This study suggests a context prediction method available for both discrete and continuous contexts without being limited to the characteristics of a specific domain or context. For this, we conducted a context prediction based on sequence matching by generating sequences from contexts in consideration of association rules between context attributes and by applying variable weights according to each context attribute. Simulations for discrete and continuous contexts were conducted to evaluate proposed methods and the results showed that the methods produced a similar performance to existing prediction methods with a prediction accuracy of 80.12% in discrete context and 81.43% in continuous context.

• The Korean Journal of Applied Statistics
• /
• v.23 no.4
• /
• pp.759-765
• /
• 2010

In this paper, two new inverse regression methods are introduced. One is a distance based method, and the other is a likelihood based method. While a model is fitted by minimizing the sum of squared prediction errors of y's and x's in the classical and inverse methods, respectively. In the new distance based method, we simultaneously minimize the sum of both squared prediction errors. In the likelihood based method, we propose an inverse regression with Arnold-Beaver Skew Normal(ABSN) error distribution. Using the cross validation method with an asymmetric real data set, two new and two existing methods are studied based on the relative prediction bias(RBP) criteria.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.10a
• /
• pp.203-206
• /
• 1996

Prediction of firm bankruptcy have been extensively studied in accounting, as all stockholders in a firm have a vested interest in monitoring its financial performance. The objective of this paper is to develop the hybrid models for bankruptcy prediction. The proposed hybrid models are two phase. Phase one are (a) DA-assisted neural network, (b) Logit-assisted neural network, and (c) Genetic-assisted neural network. And, phase two are (a) DA-assisted Case based reasoning, and (b) Genetic-assisted Case based reasoning. In the variables selection, We are focusing on three alternative methods - linear discriminant analysis, logit analysis and genetic algorithms - that can be used empirically select predictors for hybrid model in bankruptcy prediction. Empirical results using Korean medium-sized firms data show that hybrid models are very promising neural network models and case based reasoning for bankruptcy prediction in terms of predictive accuracy and adaptability.

• Journal of Broadcast Engineering
• /
• v.17 no.5
• /
• pp.781-788
• /
• 2012

In HEVC, the fast search method is used for reducing the complexity of the motion prediction procedure. It is consisted of the sub-sampled SAD which reduce the complexity of Sum of Absolute Differences(SAD) calculation and the simplified bi-prediction method which reduce the iterations of the uni-prediction for the bi-prediction. The computational complexity is largely decreased by the fast search method but the coding gain is also decreased. In this paper, the simplified bi-prediction is extended to compensate the performance loss and the prediction cost based complexity reduction methods are also proposed to reduce the complexity burden by the extended bi-prediction method. A prediction cost based complexity reduction method is consisted of early termination method for the extended bi-prediction and the bi-prediction skipping method. Compare with HM 6.0 references S/W, the average 0.42% of BD-bitrate is decreased by both the extended bi-prediction method and the prediction cost based complexity reduction methods with negligible increasement of the complexity.

• Journal of Korea Multimedia Society
• /
• v.17 no.8
• /
• pp.946-952
• /
• 2014

Mobility prediction is one of hot topics using location history information. It is useful for not only user-level applications such as people finder and recommendation sharing service but also for system-level applications such as hand-off management, resource allocation, and quality of service of wireless services. Most of current prediction techniques often use a set of significant locations without taking into account possible location information changes for prediction. Markov-based, LZ-based and Prediction by Pattern Matching techniques consider interesting locations to enhance the prediction accuracy, but they do not consider interesting location changes. In our paper, we propose an algorithm which integrates the changing or emerging new location information. This approach is based on Active LeZi algorithm, but both of new location and all possible location contexts will be updated in the tree with the fixed depth. Furthermore, the tree will also be updated even when there is no new location detected but the expected route is changed. We find that our algorithm is adaptive to predict next location. We evaluate our proposed system on a part of Dartmouth dataset consisting of 1026 users. An accuracy rate of more than 84% is achieved.