Showing 1–6 of 6 results for author: De Myttenaere, A

Mean Absolute Percentage Error for regression models

Authors: Arnaud De Myttenaere, Boris Golden, Bénédicte Le Grand, Fabrice Rossi

Abstract: We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We prove the existence of an optimal MAPE model and we show the universal consistency of Empirical Risk Minimization based on the MAPE. We also show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression,… ▽ More We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We prove the existence of an optimal MAPE model and we show the universal consistency of Empirical Risk Minimization based on the MAPE. We also show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression, and we apply this weighting strategy to kernel regression. The behavior of the MAPE kernel regression is illustrated on simulated data. △ Less

Submitted 10 July, 2017; v1 submitted 9 May, 2016; originally announced May 2016.

Journal ref: Neurocomputing, Elsevier, 2016, Advances in artificial neural networks, machine learning and computational intelligence - Selected papers from the 23rd European Symposium on Artificial Neural Networks (ESANN 2015), 192, pp.38 - 48

Study of a bias in the offline evaluation of a recommendation algorithm

Authors: Arnaud De Myttenaere, Boris Golden, Bénédicte Le Grand, Fabrice Rossi

Abstract: Recommendation systems have been integrated into the majority of large online systems to filter and rank information according to user profiles. It thus influences the way users interact with the system and, as a consequence, bias the evaluation of the performance of a recommendation algorithm computed using historical data (via offline evaluation). This paper describes this bias and discuss the r… ▽ More Recommendation systems have been integrated into the majority of large online systems to filter and rank information according to user profiles. It thus influences the way users interact with the system and, as a consequence, bias the evaluation of the performance of a recommendation algorithm computed using historical data (via offline evaluation). This paper describes this bias and discuss the relevance of a weighted offline evaluation to reduce this bias for different classes of recommendation algorithms. △ Less

Submitted 4 November, 2015; originally announced November 2015.

Comments: arXiv admin note: substantial text overlap with arXiv:1407.0822

Journal ref: Petra Perner. 11th Industrial Conference on Data Mining, ICDM 2015, Jul 2015, Hamburg, Germany. Ibai Publishing, pp.57-70, 2015, Advances in Data Mining

Empirical risk minimization is consistent with the mean absolute percentage error

Authors: Arnaud De Myttenaere, Bénédicte Le Grand, Fabrice Rossi

Abstract: We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression. We also show that, under some asumptions, universal consistency of Empirical Risk Minimization remains possible using the MAPE. We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression. We also show that, under some asumptions, universal consistency of Empirical Risk Minimization remains possible using the MAPE. △ Less

Submitted 8 September, 2015; originally announced September 2015.

Comments: in French, 47èmes Journées de Statistique de la SFdS, Jun 2015, Lille, France. 2015

Using the Mean Absolute Percentage Error for Regression Models

Authors: Arnaud De Myttenaere, Boris Golden, Bénédicte Le Grand, Fabrice Rossi

Abstract: We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression. We show that universal consistency of Empirical Risk Minimization remains possible using the MAPE instead of the MAE. We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression. We show that universal consistency of Empirical Risk Minimization remains possible using the MAPE instead of the MAE. △ Less

Submitted 12 June, 2015; originally announced June 2015.

Comments: European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN), Apr 2015, Bruges, Belgium. 2015, Proceedings of the 23-th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2015)

Reducing offline evaluation bias of collaborative filtering algorithms

Authors: Arnaud De Myttenaere, Boris Golden, Bénédicte Le Grand, Fabrice Rossi

Abstract: Recommendation systems have been integrated into the majority of large online systems to filter and rank information according to user profiles. It thus influences the way users interact with the system and, as a consequence, bias the evaluation of the performance of a recommendation algorithm computed using historical data (via offline evaluation). This paper presents a new application of a weigh… ▽ More Recommendation systems have been integrated into the majority of large online systems to filter and rank information according to user profiles. It thus influences the way users interact with the system and, as a consequence, bias the evaluation of the performance of a recommendation algorithm computed using historical data (via offline evaluation). This paper presents a new application of a weighted offline evaluation to reduce this bias for collaborative filtering algorithms. △ Less

Submitted 12 June, 2015; originally announced June 2015.

Comments: European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN), Apr 2015, Bruges, Belgium. pp.137-142, 2015, Proceedings of the 23-th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2015)

Reducing Offline Evaluation Bias in Recommendation Systems

Authors: Arnaud De Myttenaere, Bénédicte Le Grand, Boris Golden, Fabrice Rossi

Abstract: Recommendation systems have been integrated into the majority of large online systems. They tailor those systems to individual users by filtering and ranking information according to user profiles. This adaptation process influences the way users interact with the system and, as a consequence, increases the difficulty of evaluating a recommendation algorithm with historical data (via offline evalu… ▽ More Recommendation systems have been integrated into the majority of large online systems. They tailor those systems to individual users by filtering and ranking information according to user profiles. This adaptation process influences the way users interact with the system and, as a consequence, increases the difficulty of evaluating a recommendation algorithm with historical data (via offline evaluation). This paper analyses this evaluation bias and proposes a simple item weighting solution that reduces its impact. The efficiency of the proposed solution is evaluated on real world data extracted from Viadeo professional social network. △ Less

Submitted 3 July, 2014; originally announced July 2014.

Comments: 23rd annual Belgian-Dutch Conference on Machine Learning (Benelearn 2014), Bruxelles : Belgium (2014)