Evaluating Demand Prediction Techniques for Computational Markets
نویسنده
چکیده
We evaluate different prediction techniques to estimate future demand of resource usage in a computational market. Usage traces from the PlanetLab network are used to compare the prediction accuracy of models based on histograms, normal distribution approximation, maximum entropy, and autoregression theory. We particularly study the ability to predict the tail of the probability distribution in order to give guarantees of upper bounds of demand. We found that the maximum entropy model was particularly well suited to predict these upper bounds.
منابع مشابه
Statistical Methods for Computational Markets Proportional Share Market Prediction and Admission Control
We design, implement and evaluate statistical methods for managing uncertainty when consuming and provisioning resources in a federated computational market. To enable efficient allocation of resources in this environment, providers need to know consumers’ risk preferences, and the expected future demand. The guarantee levels to offer thus depend on techniques to forecast future usage and to ac...
متن کاملPrice Elasticity in Electronic Markets: Evaluating Quality and Product Information for Search and Experience Products
Evidence has shown that the provision of product information in electronic markets decreases the price elasticity of demand due to the ‘fit’ cost. This effect, however, could differ according to how consumers perceive the value of the product information to their quality evaluation procedures. If the information has very limited value, then they may not rely on it; thus, the demand elasticity m...
متن کاملWind Power Prediction with Machine Learning Ensembles
For a sustainable integration of wind power into the electricity grid, precise and robust predictions are required. With increasing installed capacity and changing energy markets, there is a growing demand for short-term predictions. Machine learning methods can be used as a purely data-driven, spatio-temporal prediction model that yields better results than traditional physical models based on...
متن کاملFostering efficiency of computational resource allocation - Integrating information services into markets
The application of market mechanisms for the allocation of computing services is a demanding task, which requires bridging economic and associated technical challenges. Even if the market-based approach promises an efficient allocation of computing services, the wide heterogeneity of consumer requirements and the diversity of computational services on provider side are challenging the processes...
متن کاملQuantitative Estimation of Market Sentiment: a discussion of two alternatives
Modeling and estimating the time series (the value in time) of assets traded in real financial markets is an intriguing challenge that has attracted researchers from many fields including Economics, Statistics, and more recently Computer Science thanks to the common availability of computational power that made possible the investigation of computational or simulative methods for modeling finan...
متن کامل