• No results found

Bioinformatics Engineering Program Uppsala University School of Engineering

N/A
N/A
Protected

Academic year: 2022

Share "Bioinformatics Engineering Program Uppsala University School of Engineering"

Copied!
1
0
0

Loading.... (view fulltext now)

Full text

(1)

Bioinformatics Engineering Program

Uppsala University School of Engineering

UPTEC X 09 026 Date of issue 2009-06

Author

Christofer Bäcklin

Title (English)

Performance estimation of multivariate classification of blood samples for cancer detection

Title (Swedish)

Prestandautvärdering vid multivariat klassning av blodprover för cancerdetektion

Abstract

Disease detection methods involving multivariate classification need reliable performance estimates to be practically useful. A new method for making such estimates is presented here based on splitting the data in several bags and estimating the error rate by training many classifiers on new data drawn from kernel density estimations of the unknown underlying distribution in each bag. This gives an estimation of the entire prior function of the classification problem together with an uncertainty measure which is a great improvement from the standard methods of today which only supply its mean and no uncertainty measure.

Keywords

Classification performance, error rate, kernel estimation, cancer detection, in solution PLA Supervisors

Claes Andersson

Uppsala university, department of medical sciences

Scientific reviewer

Mats Gustafsson

Uppsala universitet, the Linnaeus center for bioinformatics

Project name Sponsors

Language

Swedish Security Secret until 18

th

June 2011

ISSN 1401-2138 Classification

Supplementary bibliographical information

Pages

27

Biology Education Centre Biomedical Center Husargatan 3 Uppsala

Box 592 S-75124 Uppsala Tel +46 (0)18 4710000 Fax +46 (0)18 555217

References

Related documents

This degree project involved the implementation of various reports to study the dose-response relationship, as well as preparing a general system for data extraction to 3rd

Multiplex ligation dependent genome amplification (MLGA) is a novel method for the multiplex targeted measurement and validation of genetic copy-number variations (CNVs). A

Datex II is a European standard used for exchange of traffic data that the Swedish Road Administration (SRA) uses to supply its service providers with relevant information

This thesis work investigates the effects of microsomal Prostaglandin E Synthase-1 (mPGES- 1) inhibition on the functional activity of synovial fibroblasts from patients with

The aim of this thesis is to develop a process for the selection of a target for the identification of microorganisms using the so-called padlock probes, and to design and implement a

In this study the performance of four classification algorithms; k-Nearest Neighbors, Linear Discriminant Analysis, Support Vector Machine and Random Forest was analysed when

Differential expression, mRNA, R, Bowtie, adenovirus, Solexa, time series, data analysis, human

A text extraction tool was implemented on the .NET platform with functionality for preprocessing text (removal of stop words, Porter stemming and use of synonyms) and matching