Constructing Ensembles from Data Envelopment Analysis

نویسندگان

  • Zhiqiang Zheng
  • Balaji Padmanabhan
چکیده

It has been shown in prior work in management science, statistics and machine learning that using an ensemble of models often results in better performance than using a single ‘best’ model. This paper proposes a novel Data Envelopment Analysis (DEA) based approach to combine models. We prove that for the 2-class classification problems, DEA models identify the same convex hull as the popular ROC analysis used for model combination. We further develop two DEA-based methods to combine k-class classifiers. Experiments demonstrate that the two methods outperform other benchmark methods and suggest that DEA can be a powerful tool for model combination. ______________________________ A preliminary version of this paper was accepted at the ACM Conference on Knowledge Discovery and Data Mining 2004 (KDD04). This version substantially extends the conference publication by a comprehensive literature review, a better model combination method, theoretical proofs and more

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

DATA ENVELOPMENT ANALYSIS WITH FUZZY RANDOM INPUTS AND OUTPUTS: A CHANCE-CONSTRAINED PROGRAMMING APPROACH

In this paper, we deal with fuzzy random variables for inputs andoutputs in Data Envelopment Analysis (DEA). These variables are considered as fuzzyrandom flat LR numbers with known distribution. The problem is to find a method forconverting the imprecise chance-constrained DEA model into a crisp one. This can bedone by first, defuzzification of imprecise probability by constructing a suitablem...

متن کامل

Supply Chain Performance Evaluation Using Data Envelopment Analysis (A case study of tile industry)

The competition increscent urgency in the global economy causes many organizations focus more on novel management ideas in constructing, delivery services and their productions. In current conditions, supply chain and the importance of a suitable performance evaluating system in order to understand the current condition and program for its improvement, has a very special importance in creating ...

متن کامل

Portfolio Performance Evaluation in a Modified Mean-Variance-Skewness Framework with Negative Data

   The present study is an attempt toward evaluating the performance of portfolios using mean-variance-skewness model with negative data. Mean-variance non-linear framework and mean-variance-skewness non- linear framework had been proposed based on Data Envelopment Analysis, which the variance of the assets had been used as an input to the DEA and expected return and skewness were the output. C...

متن کامل

An Evaluation of Factory Performance Utilized Kpi/kai with Data Envelopment Analysis

In this paper, based on Key Performance Indicator (KPI) and Key Activity Indicator (KAI) as the fundamental data for calculating defined factory performance function, evaluation system of factory performance is discussed by following four phases. 1) Designing framework for evaluation system of factory performance, 2) Constructing structure of KPI/KAI database, 3) Formulating evaluation of facto...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:
  • INFORMS Journal on Computing

دوره 19  شماره 

صفحات  -

تاریخ انتشار 2007