Well-Calibrated Predictions from Online Compression Models

نویسنده

  • Vladimir Vovk
چکیده

It has been shown recently that Transductive Confidence Machine (TCM) is automatically well-calibrated when used in the on-line mode and provided that the data sequence is generated by an exchangeable distribution. In this paper we strengthen this result by relaxing the assumption of exchangeability of the data-generating distribution to the much weaker assumption that the data agrees with a given “on-line compression model”.

برای دانلود رایگان 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

Calibrated Confidence Measures With Guarantees

In medical and healthcare applications of machine learning, assessing the confidence of predictions is often as important as obtaining high accuracy. Here, we present an algorithm for constructing calibrated confidence estimates, which are probabilities that match empirical event frequencies in the long run. Unlike previous approaches, our confidence scores are generated in an online fashion an...

متن کامل

Determining the best water quality model for the rivers in north of Iran (case study: Pasikhan River)

The aim of this paper is to study the popular models used for the simulation of major elements of a water quality system. The present study compares results of the CE-QUAL-W2 model with the WASP5 and MIKE11 models to assess the water quality of Pasikhan River. The contaminant loadings of Nitrate and Phosphate are utilized in the CE-QUAL-W2, WASP5 and MIKE11 simulations. This study is the first ...

متن کامل

A Universal Well-Calibrated Algorithm for On-line Classification

We study the problem of on-line classification in which the prediction algorithm, for each “significance level” δ, is required to output as its prediction a range of labels (intuitively, those labels deemed compatible with the available data at the level δ) rather than just one label; as usual, the examples are assumed to be generated independently from the same probability distribution P. The ...

متن کامل

Analytical predictions for the buckling of a nanoplate subjected to non-uniform compression based on the four-variable plate theory

In the present study, the buckling analysis of the rectangular nanoplate under biaxial non-uniform compression using the modified couple stress continuum theory with various boundary conditions has been considered. The simplified first order shear deformation theory (S-FSDT) has been employed and the governing differential equations have been obtained using the Hamilton’s principle. An analytic...

متن کامل

Validation and calibration of the Kabi Pharmacia International Growth Study prediction model for children with idiopathic growth hormone deficiency.

In 1999 a model was published for prediction of growth in children with idiopathic GH deficiency (IGHD) during GH therapy, derived using data from the Kabi Pharmacia International Growth Study (KIGS) database (Pharmacia \|[amp ]\| Upjohn, Inc., International Growth Database). We validated and calibrated this KIGS model for growth in the first year of GH therapy using data from 136 Dutch childre...

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

عنوان ژورنال:
  • Theor. Comput. Sci.

دوره 364  شماره 

صفحات  -

تاریخ انتشار 2003