Statistical Evaluation of Complex Input-Output Transformations

Author

Shu, Gang ; Bai, Zhuofu ; Podgurski, Andy

Author_Institution

EECS Dept., Case Western Reserve Univ., Cleveland, OH, USA

fYear

2011

fDate

Nov. 29 2011-Dec. 2 2011

Firstpage

70

Lastpage

79

Abstract

This paper presents a new, statistical approach to evaluating software products that transform complex inputs into complex outputs. This approach, called multistage stratified input/output (MSIO) sampling, combines automatic clustering of multidimensional I/O data with multistage sampling and manual examination of data elements, in order to accurately and economically estimate summary measures of output data quality. We report results of two case studies in which MSIO sampling was successfully applied to evaluating complex graphical outputs.

Keywords

data analysis; input-output programs; pattern clustering; program testing; sampling methods; software metrics; software reliability; MSIO sampling; complex graphical outputs evaluation; complex input-output transformations; data element manual examination; multidimensional I/O data automatic clustering; multistage stratified input-output sampling; output data quality; software product evaluation; software reliability; software testing; statistical approach; statistical evaluation; summary measure estimation; Estimation; Handheld computers; Java; Software; Software measurement; Software reliability; Testing; MSIO sampling; bioinformatics; complex data; estimation; multistage sampling; software reliability; software testing; stratified sampling;

fLanguage

English

Publisher

ieee

Conference_Titel

Software Reliability Engineering (ISSRE), 2011 IEEE 22nd International Symposium on

Conference_Location

Hiroshima

ISSN

1071-9458

Print_ISBN

978-1-4577-2060-4

Type

conf

DOI

10.1109/ISSRE.2011.32

Filename

6132955