Comparison of Methods for Detecting Outliers in Medical Data

الموضوعات

إحصاء

الملخص

Observed variables often contain outliers that have unusually large or small values when compared with others in a data set. Some data sets may come from homogeneous groups; others from heterogeneous groups. Outliers can be caused by incorrect measurements, including data entry errors, or by coming from a different population than the rest of the data. The Appearance of the outlier Values in the set of data may effect on the result of the statistical Analysis of data, from this point of view appears the Importance of detecting the outlier values in any set of data before study. To detect the mentioned outliers, we have used formal and informal methods. Whatever, this study aimed at making a comparison among the detection outliers methods (formal and informal methods) to discover the best in detecting outliers. To achieve this goal, we have simulated 4 samples with different size which follow the normal distribution N(0,1). The sizes of these samples are: 20, 100, 500 and 1000. Another elements which follow the normal distribution N(0,1.5) had been added to the previous samples with a percentage of 20% for each one. The new sizes became 24, 120, 600 and 1200. After applying these methods (formal and informal) on the samples to detect the outliers, we have noticed that: Tukey which is an informal method had the highest sensitivity in detecting outliers, next one is MAD and then the modified Z-score. Z and SD methods were of low sensitivity in detecting outliers. For formal methods, Quartile and Hample methods were the most effective in detecting outliers, next one is ESD. Dixon is not appropriate in detecting outliers in samples more than 30 elements. Grubbs was of low sensitivity in detecting outliers. We have applied the mentioned methods on a medical data set of body mass index (BMI) with a sample of 309, this sample was of patients of UN sanitation center in Khanyounis, their ages were ranging between 25- 50 years old. Results were the same of simulation study as Hample and Quartile as formal methods were the most effective in detecting outliers and Tukey as informal method was the most effective then MAD.

روابط وملفات

التعريف والنوع

رقم الوثيقة

0d3a34b2-e364-481b-93f7-ce934db5ab0c

رقم العقد

0

نوع الوسائط

Crawler

نوع المحتوى

الرسائل العلمية

صيغة المصدر

رسائل ماجيستير

نوع الملف

pdf text

أسماء الملفات

2103591_1.pdf

بيانات النشر

ألقاب المؤلفين

[{"name_ar":"Abdalla M. EL-Habil","title_ar":"اشراف","title_en":"Supervision"},{"name_ar":"Jannat Abdallah AL-Astal","title_ar":"اعداد","title_en":"Preparation"}]

اللغة

English

المصدر والدورية

اسم المصدر

Comparison of Methods for Detecting Outliers in Medical Data

المحتوى والصفحات

عدد الصفحات

0

إشراف وإعداد

الإشراف

Abdalla M. EL-Habil

الإعداد

Jannat Abdallah AL-Astal

الاقتباسات الببليوغرافية