The Experts below are selected from a list of 9432 Experts worldwide ranked by ideXlab platform
Peter Bühlmann - One of the best experts on this subject based on the ideXlab platform.
-
Tree‐structured generalized autoregressive conditional heteroscedastic models
Journal of the Royal Statistical Society: Series B (Statistical Methodology), 2001Co-Authors: Francesco Audrino, Peter BühlmannAbstract:We propose a new generalized autoregressive conditional heteroscedastic (GARCH) model with tree-structured multiple thresholds for the estimation of volatility in financial time series. The approach relies on the idea of a binary tree where every Terminal Node parameterizes a (local) GARCH model for a partition cell of the predictor space. The fitting of such trees is constructed within the likelihood framework for non-Gaussian observations: it is very different from the well-known regression tree procedure which is based on residual sums of squares. Our strategy includes the classical GARCH model as a special case and allows us to increase model complexity in a systematic and flexible way. We derive a consistency result and conclude from simulation and real data analysis that the new method has better predictive potential than other approaches.
-
Tree structured GARCH models
2000Co-Authors: Francesco Audrino, Peter BühlmannAbstract:We propose a new GARCH model with tree-structured multiple thresholds for volatility estimation in financial time series. The approach relies on the idea of a binary tree where every Terminal Node parameterizes a (local) GARCH model for a partition cell of the predictor space. Fitting of such trees is constructed within the likelihood framework for non-Gaussian observations: it is very different from the well-known CART procedure for regression based on residual sum of squares. Our strategy includes the classical GARCH model and allows in a systematic and flexible way to increase model-complexity. We conclude with simulations and real data analysis that the new method has better predictive potential compared to other approaches.
Zhang Guo-yin - One of the best experts on this subject based on the ideXlab platform.
-
Concurrent Transmission MAC Protocol for Wireless Sensor Network Based on Nodes' Geographical Location Information
Computer Science, 2012Co-Authors: Zhang Guo-yinAbstract:Aiming at the exposed Terminal problems in multi-hop wireless sensor networks,this paper proposed the efficient concurrent transmission LACT-MAC protocol based on geographical location information.The protocol breaks through the limit of concurrent transmission based on traditional CSMA MAC protocols,for this protocol exploits Nodes' geographical location information to achieve the concurrent transmission of exposed Terminal,so as to enhance the reuse efficiency of the valuable wireless channel resources.According to the location coordinates of the Nodes,this paper explored the feasibility analysis of concurrent transmission of exposed Terminal Node,and completed the concurrency check.Simulation results show that,compared to IEEE 802.11 DCF protocol,LACT-MAC protocol can significantly improve the average throughput of the network,reduce data transmission delay,and effectively improve the efficiency and performance in wireless sensor networks.
Francesco Audrino - One of the best experts on this subject based on the ideXlab platform.
-
Tree‐structured generalized autoregressive conditional heteroscedastic models
Journal of the Royal Statistical Society: Series B (Statistical Methodology), 2001Co-Authors: Francesco Audrino, Peter BühlmannAbstract:We propose a new generalized autoregressive conditional heteroscedastic (GARCH) model with tree-structured multiple thresholds for the estimation of volatility in financial time series. The approach relies on the idea of a binary tree where every Terminal Node parameterizes a (local) GARCH model for a partition cell of the predictor space. The fitting of such trees is constructed within the likelihood framework for non-Gaussian observations: it is very different from the well-known regression tree procedure which is based on residual sums of squares. Our strategy includes the classical GARCH model as a special case and allows us to increase model complexity in a systematic and flexible way. We derive a consistency result and conclude from simulation and real data analysis that the new method has better predictive potential than other approaches.
-
Tree structured GARCH models
2000Co-Authors: Francesco Audrino, Peter BühlmannAbstract:We propose a new GARCH model with tree-structured multiple thresholds for volatility estimation in financial time series. The approach relies on the idea of a binary tree where every Terminal Node parameterizes a (local) GARCH model for a partition cell of the predictor space. Fitting of such trees is constructed within the likelihood framework for non-Gaussian observations: it is very different from the well-known CART procedure for regression based on residual sum of squares. Our strategy includes the classical GARCH model and allows in a systematic and flexible way to increase model-complexity. We conclude with simulations and real data analysis that the new method has better predictive potential compared to other approaches.
G. V. Conroy - One of the best experts on this subject based on the ideXlab platform.
-
Interval error estimators in class probability trees
Pattern Recognition Letters, 1996Co-Authors: G. Kalkanis, G. V. ConroyAbstract:This paper suggests the use of confidence interval estimators in order to control the design of a class probability tree from a set of past data. Class probability trees are decision trees in which each Terminal Node is assigned an estimate of the probabilities of all the decision classes.
Rakesh Kawatra - One of the best experts on this subject based on the ideXlab platform.
-
A multiperiod min-sum arborescence problem
OPSEARCH, 2014Co-Authors: Rakesh KawatraAbstract:In this paper we present a mathematical formulation of the multiperiod min-sum arborescence problem which requires scheduling the installation of directed links to connect a set of Terminal Nodes N = {2,3,…,n} to a central Node. The selection of links and the timing of their installation should be such that the present value of expenditures including cost of installing all the links and maintaining them till the end of the planning horizon should be minimal. The set of links selected for installation should ensure that (i) each Terminal Node j has exactly one entering link; and (ii) for each Terminal Node j, a unique path exists from the central Node to Node j from the period when j installed till the end of the planning horizon. Some of the Terminal Nodes are included in the network at the beginning of the planning horizon while others are added to the network over time. We present a branch exchange heuristic embedded in the Lagrangian relaxation method to find a solution to this problem. The lower bound given by our solution method is used to estimate the quality of the solution given by the heuristic. Test results over a wide range of problem structures show that for networks with up to 140 Nodes our Lagrangian-based heuristic method gives solutions that are within 5–15 % of optimality.
-
A hop constrained min-sum arborescence with outage costs
Computers & Operations Research, 2007Co-Authors: Rakesh KawatraAbstract:The hop constrained min-sum arborescence with outage costs problem consists of selecting links in a network so as to connect a set of Terminal Nodes N={2,3,...,n} to a central Node with minimal expected annual cost. The maximum number of links between the central Node and each Terminal Node j is limited to a predefined number h"j (the hop constraint). Each Terminal Node in the network has an associated outage cost, which is the economic cost borne by the network user whenever that Node is disconnected from the central Node due to failure of a link. We formulate this problem as an integer programming problem and suggest a Lagrangian relaxation-based heuristic to get a good solution to this network problem. Lower bounds found as a byproduct of the solution procedure are used to assess the quality of the heuristic solutions. Computational results over a wide range of problem structures involving up to 100 Nodes are given showing the effectiveness of the proposed approach.
-
HICSS - A hop constrained min-sum arborescence with outage costs
36th Annual Hawaii International Conference on System Sciences 2003. Proceedings of the, 2003Co-Authors: Rakesh KawatraAbstract:The hop constrained min-sum arborescence with outage costs problem consists of selecting links in a network so as to connect a set of Terminal Nodes N={2,3, ...n} to a central Node with minimal total link cost such that (a) each Terminal Node j has exactly one entering link; (b) for each Terminal Node j, a unique path from the central Node to j exists; (c) for each Terminal Node j the number of links between the central Node and j is limited to a predefined number h/sub j/, and (d) each Terminal Node has an associated outage cost, which is the economic cost incurred by the network user whenever that Node is disabled due to failure of a link. We suggest a Lagrangian based heuristic to solve the integer programming formulation of this network problem.
-
Lower Bounds for the Multiperiod Capacitated Minimal Spanning Tree with Node Outage Cost Design Problem
2000Co-Authors: Rakesh KawatraAbstract:The Multiperiod Capacitated Minimal Spanning Tree With Node Outage Costs (MCMSTWOC) Design problem consists of scheduling the installation of links in a communication network so as to connect a set of Terminal Nodes S = [2,3...N] to a central Node (Node 1) with minimal present value of costs. The cost of the network is the sum of link layout cost and Node outage costs. The link capacities limit the number of Terminal Nodes sharing a link. Node outage cost associated with each Terminal Node is the economic cost incurred by the network user whenever the Terminal Node is disabled due to failure of a link. In the network some of the Terminal Nodes are active at the beginning of the planning horizon while others are activated over time. The problem is formulated as an integer-programming problem. A Lagrangian relaxation method is used to find a lower bound for the optimal objective function value. Subgradient optimization method is used to find good lower bounds. This lower bound can be used to estimate the quality of the solution given by a heuristic.
