Results 11  20
of
200
Crossvalidation and the bootstrap: Estimating the error rate of a prediction rule
, 1995
"... ..."
(Show Context)
Bayesian Model Assessment and Comparison Using CrossValidation Predictive Densities
 Neural Computation
, 2002
"... In this work, we discuss practical methods for the assessment, comparison, and selection of complex hierarchical Bayesian models. A natural way to assess the goodness of the model is to estimate its future predictive capability by estimating expected utilities. Instead of just making a point estimat ..."
Abstract

Cited by 47 (16 self)
 Add to MetaCart
(Show Context)
In this work, we discuss practical methods for the assessment, comparison, and selection of complex hierarchical Bayesian models. A natural way to assess the goodness of the model is to estimate its future predictive capability by estimating expected utilities. Instead of just making a point estimate, it is important to obtain the distribution of the expected utility estimate, as it describes the uncertainty in the estimate. The distributions of the expected utility estimates can also be used to compare models, for example, by computing the probability of one model having a better expected utility than some other model. We propose an approach using crossvalidation predictive densities to obtain expected utility estimates and Bayesian bootstrap to obtain samples from their distributions. We also discuss the probabilistic assumptions made and properties of two practical crossvalidation methods, importance sampling and kfold crossvalidation. As illustrative examples, we use MLP neural networks and Gaussian Processes (GP) with Markov chain Monte Carlo sampling in one toy problem and two challenging realworld problems.
Forecast combination and model averaging using predictive measures’, Econometric Reviews 26
, 2007
"... We extend the standard approach to Bayesian forecast combination by forming the weights for the forecast combinations from the predictive likelihood rather than the standard marginal likelihood. The use of predictive measures of fit offer greater protection against insample overfitting and improve ..."
Abstract

Cited by 38 (0 self)
 Add to MetaCart
We extend the standard approach to Bayesian forecast combination by forming the weights for the forecast combinations from the predictive likelihood rather than the standard marginal likelihood. The use of predictive measures of fit offer greater protection against insample overfitting and improves forecast performance. For the predictive likelihood we show analytically that the forecast weights have good large and small sample properties. This is confirmed in a simulation study and an application to forecasts of the Swedish inflation rate where forecast combination using the predictive likelihood outperforms standard Bayesian model averaging using the marginal likelihood.
The information content of the forwardlooking statements in corporate filings  A Naive Bayesian machine learning approach
 Journal of Accounting Research
"... This paper examines the tone and content of the forwardlooking statements (FLS) in the Management Discussion and Analysis section (MD&A) of corporate 10K and 10Q filings using a Naïve Bayesian machine learning algorithm. I first manually categorize 30,000 sentences of randomlyselected FLS ex ..."
Abstract

Cited by 34 (1 self)
 Add to MetaCart
This paper examines the tone and content of the forwardlooking statements (FLS) in the Management Discussion and Analysis section (MD&A) of corporate 10K and 10Q filings using a Naïve Bayesian machine learning algorithm. I first manually categorize 30,000 sentences of randomlyselected FLS extracted from the MD&As along two dimensions: (1) tone (i.e., positive versus negative tone); and (2) content (i.e., profitability, operations, liquidity etc.). These manuallycoded sentences are then used as training data in a Naïve Bayesian machine learning algorithm to classify the tone and content of about 13 million forwardlooking statements from more than 140,000 corporate 10K and 10Q MD&As between 1994 and 2007. I find that firms with better current performance, lower accruals, smaller size, lower markettobook ratio, and less return volatility tend to have more positive forwardlooking
Predicting the Stock Market
, 1998
"... This paper presents a tuturial introduction to predictions of stock time series. The various approaches of technical and fundamental analysis is presented and the prediction problem is formulated as a special case of inductive learning. The problems with performance evaluation of nearrandomwalk pr ..."
Abstract

Cited by 31 (1 self)
 Add to MetaCart
(Show Context)
This paper presents a tuturial introduction to predictions of stock time series. The various approaches of technical and fundamental analysis is presented and the prediction problem is formulated as a special case of inductive learning. The problems with performance evaluation of nearrandomwalk processes are illustrated with examples together with guidelines for avoiding the risk of datasnooping. The connections to concepts like "the biasvariance dilemma", overtraining and model complexity are further covered. Existing benchmarks and testing metrics are surveyed and some new measures are introduced.
Crosssystem user modeling and personalization on the social web. UMUAI  Journal of User Modeling and UserAdapted Interaction, forthcoming
"... Abstract. In order to adapt functionality to their individual users, systems need information about these users. The Social Web provides opportunities to gather user data from outside the system itself. Aggregated user data may be useful to address coldstart problems as well as sparse user profiles ..."
Abstract

Cited by 30 (6 self)
 Add to MetaCart
(Show Context)
Abstract. In order to adapt functionality to their individual users, systems need information about these users. The Social Web provides opportunities to gather user data from outside the system itself. Aggregated user data may be useful to address coldstart problems as well as sparse user profiles, but this depends on the nature of individual user profiles distributed on the Social Web. For example, does it make sense to reuse Flickr profiles to recommend bookmarks in Delicious? In this article, we study distributed formbased and tagbased user profiles, based on a large dataset aggregated from the Social Web. We analyze the completeness, consistency and replication of formbased profiles, which users explicitly create by filling out forms at Social Web systems such as Twitter, Facebook and LinkedIn. We also investigate tagbased profiles, which result from social tagging activities in systems such as Flickr, Delicious and StumbleUpon: to what extent do tagbased profiles overlap between different systems, what are the benefits of aggregating tagbased profiles. Based on these insights, we developed and evaluated the performance of several crosssystem user modeling strategies in the context of recommender systems. The evaluation results show that the proposed methods solve the coldstart problem and improve recommendation quality significantly, even beyond the coldstart. 1
Bayesian Approach for Neural Networks  Review and Case Studies
 Neural Networks
, 2001
"... We give a short review on the Bayesian approach for neural network learning and demonstrate the advantages of the approach in three real applications. We discuss the Bayesian approach with emphasis on the role of prior knowledge in Bayesian models and in classical error minimization approaches. The ..."
Abstract

Cited by 28 (10 self)
 Add to MetaCart
(Show Context)
We give a short review on the Bayesian approach for neural network learning and demonstrate the advantages of the approach in three real applications. We discuss the Bayesian approach with emphasis on the role of prior knowledge in Bayesian models and in classical error minimization approaches. The generalization capability of a statistical model, classical or Bayesian, is ultimately based on the prior assumptions. The Bayesian approach permits propagation of uncertainty in quantities which are unknown to other assumptions in the model, which may be more generally valid or easier to guess in the problem. The case problems studied in this paper include a regression, a classification, and an inverse problem. In the most thoroughly analyzed regression problem, the best models were those with less restrictive priors. This emphasizes the major advantage of the Bayesian approach, that we are not forced to guess attributes that are unknown, such as the number of degrees of freedom in the model, nonlinearity of the model with respect to each input variable, or the exact form for the distribution of the model residuals.
Estimating the genetic architecture of quantitative traits
, 1999
"... Understanding and estimating the structure and parameters associated with the genetic architecture of quantitative traits is a major research focus in quantitative genetics. With the availability of a wellsaturated genetic map of molecular markers, it is possible to identify a major part of the str ..."
Abstract

Cited by 28 (4 self)
 Add to MetaCart
(Show Context)
Understanding and estimating the structure and parameters associated with the genetic architecture of quantitative traits is a major research focus in quantitative genetics. With the availability of a wellsaturated genetic map of molecular markers, it is possible to identify a major part of the structure of the genetic architecture of quantitative traits and to estimate the associated parameters. Multiple interval mapping, which was recently proposed for simultaneously mapping multiple quantitative trait loci (QTL), is well suited to the identification and estimation of the genetic architecture parameters, including the number, genomic positions, effects and interactions of significant QTL and their contribution to the genetic variance. With multiple traits and multiple environments involved in a QTL mapping experiment, pleiotropic effects and QTL by environment interactions can also be estimated. We review the method and discuss issues associated with multiple interval mapping, such as likelihood analysis, model selection, stopping rules and parameter estimation. The potential power and advantages of the method for mapping multiple QTL and estimating the genetic architecture are discussed. We also point out potential problems and difficulties in resolving the details of the genetic architecture as well as other areas that require further investigation. One application of the analysis is to improve genomewide markerassisted selection, particularly when the information about epistasis is used for selection with mating. 1.
Alliance entrepreneurship and firm market performance
 Strategic Management Journal
, 2001
"... This paper extends entrepreneurship into the domain of alliances, and investigates the effect of alliance proactiveness on marketbased firm performance (defined in terms of sales growth, market share, market development and product development). Alliance proactiveness is defined as the extent to wh ..."
Abstract

Cited by 26 (0 self)
 Add to MetaCart
(Show Context)
This paper extends entrepreneurship into the domain of alliances, and investigates the effect of alliance proactiveness on marketbased firm performance (defined in terms of sales growth, market share, market development and product development). Alliance proactiveness is defined as the extent to which an organization engages in identifying and responding to partnering opportunities. The effect of alliance proactiveness on performance is tested within a contingency framework, with size and perceived environmental uncertainty as moderators, and using data from 182 firms. We estimated the model using partial least squares. Results indicate that alliance proactiveness leads to superior marketbased performance, and that this effect is stronger for small firms and in unstable market environments. Copyright © 2001 John Wiley & Sons, Ltd. Entrepreneurship, which typically leads to new product introduction or market entry, creates value through association with the discovery and exploitation of profitable business opportunities, (Shane and Venkataraman, 2000; Lumpkin and Dess, 1996). In addition, entrepreneurial activities also create value when they facilitate ‘access relationships ’ to resources and capabilities that are strategic to competitiveness and performance (Stuart, 2000). Consequently, although extant literature has focused predominantly on entrepreneurship in product markets, entrepreneurial opportunities also exist in factor markets (Shane and Venkataraman, 2000; Schumpeter, 1934). Strategic factor markets have been defined as ‘market(s) where the resources necessary to implement a strategy are acquired ’ (Barney, 1986: 1231). For instance, the relevant strategic factor Key words: strategic alliances; entrepreneurship; proactiveness; partial least squares