Font Size: a A A

Research On Quality Evaluation Methods For Big Data Application Software

Posted on:2019-12-04Degree:MasterType:Thesis
Country:ChinaCandidate:H ChenFull Text:PDF
GTID:2438330572462848Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the development of big data techonology,big data applications have been widely used in daily life.However,the quality of big data software is not perfect.There are quality defects in big data software such as usage quality and consumer satisfaction,and those quality defects have brought huge economic losses.Due to the lack of quality evaluation standard methods for big data software,this paper focuses on the quality evaluation method of big data systems.We analyze the loss brought by big data application quality problems and existing related work on big data application quality evaluation.We propose quality models for four popular big data application systems,and refine the quality factors and generate the computing measures for search engines and age recognition systems.Through experiment we have found the quality defects and the effectiveness of the quality factors and measures is verified.(1)We study on 4 typical big data systems(search engine,image recognition,prediction system,recommendation system),and propose the quality model and express quality factors and measures using fish bone pictures.(2)Aimed at the search engines in shopping platform,we optimize the quality model and propose 6 quality factors and the corresponding 17 quality measures,including 3 computing measures for relativity,2 measures for data stability and 3 for rank stability,4 measures for intelligent processing,4 measures for integrity and 1 for rank science degree.We apply the 17 measures to 6 shopping platforms and compare the results.We have found some quality defects and verified the effectiveness of the quality factors and computing measures.(3)Aimed at the face age system,we optimize the quality modle and propose 4 quality factors and the corresponding 12 quality measures for it using error analyzing and metamorphic testing,including 3 computing measures for recognition rate,3 measures for accurancy,5 metamorphic relations for robustness and 1 measures for response time.We apply the 12 measures to face age recognition systems and compare the results.We have found some quality defects.
Keywords/Search Tags:big data application software, quality defect, quality evaluation, quality model, quality factor, computing measures
PDF Full Text Request
Related items