دورية أكاديمية
Tools to identify linear combination of prognostic factors which maximizes area under receiver operator curve.
العنوان: | Tools to identify linear combination of prognostic factors which maximizes area under receiver operator curve. |
---|---|
المؤلفون: | Todor N; Oncology Institute 'Prof.Dr. Ion Chiricuta', Biostatistics and Medical Informatics Department, Republicii 34-36, 400015 Cluj-Napoca, Romania., Todor I; Oncology Institute 'Prof.Dr. Ion Chiricuta', Radiotherapy Department, Republicii 34-36, 400015 Cluj-Napoca, Romania., Săplăcan G; Company for Applied Informatics, Republicii 101-102, 400015 Cluj-Napoca, Romania. |
المصدر: | Journal of clinical bioinformatics [J Clin Bioinforma] 2014 Jul 04; Vol. 4, pp. 10. Date of Electronic Publication: 2014 Jul 04 (Print Publication: 2014). |
نوع المنشور: | Journal Article |
اللغة: | English |
بيانات الدورية: | Publisher: Biomed Central Ltd Country of Publication: England NLM ID: 101554637 Publication Model: eCollection Cited Medium: Print ISSN: 2043-9113 (Print) Linking ISSN: 20439113 NLM ISO Abbreviation: J Clin Bioinforma Subsets: PubMed not MEDLINE |
أسماء مطبوعة: | Original Publication: [London, U. K.] : Biomed Central Ltd., c2011- |
مستخلص: | Background: The linear combination of variables is an attractive method in many medical analyses targeting a score to classify patients. In the case of ROC curves the most popular problem is to identify the linear combination which maximizes area under curve (AUC). This problem is complete closed when normality assumptions are met. With no assumption of normality search algorithm are avoided because it is accepted that we have to evaluate AUC n(d) times where n is the number of distinct observation and d is the number of variables. Methods: For d = 2, using particularities of AUC formula, we described an algorithm which lowered the number of evaluations of AUC from n(2) to n(n-1) + 1. For d > 2 our proposed solution is an approximate method by considering equidistant points on the unit sphere in R(d) where we evaluate AUC. Results: The algorithms were applied to data from our lab to predict response of treatment by a set of molecular markers in cervical cancers patients. In order to evaluate the strength of our algorithms a simulation was added. Conclusions: In the case of no normality presented algorithms are feasible. For many variables computation time could be increased but acceptable. |
References: | Biostatistics. 2000 Jun;1(2):123-40. (PMID: 12933515) Stat Med. 2011 Jul 20;30(16):2005-14. (PMID: 21472763) Biometrics. 2007 Sep;63(3):751-7. (PMID: 17825007) Stat Med. 1998 May 15;17(9):1033-53. (PMID: 9612889) Am J Epidemiol. 2004 May 1;159(9):882-90. (PMID: 15105181) Med Decis Making. 2008 Nov-Dec;28(6):909-16. (PMID: 18556633) Biostatistics. 2003 Jul;4(3):341-51. (PMID: 12925502) Biometrics. 2000 Jun;56(2):352-9. (PMID: 10877289) Rom J Morphol Embryol. 2011;52(1):53-9. (PMID: 21424032) Biometrics. 1998 Mar;54(1):124-35. (PMID: 9544511) Med Decis Making. 2004 Nov-Dec;24(6):659-69. (PMID: 15534346) |
فهرسة مساهمة: | Keywords: Area under curve; Linear combination; Receiver operator characteristics; Sensitivity; Specificity |
تواريخ الأحداث: | Date Created: 20140729 Date Completed: 20140728 Latest Revision: 20211021 |
رمز التحديث: | 20240628 |
مُعرف محوري في PubMed: | PMC4099021 |
DOI: | 10.1186/2043-9113-4-10 |
PMID: | 25068036 |
قاعدة البيانات: | MEDLINE |
تدمد: | 2043-9113 |
---|---|
DOI: | 10.1186/2043-9113-4-10 |