Açık Uçlu Gruplanmış Veriler için Ortalama ve Standart Sapma
Verilerin daha iyi açıklanması ve hesaplamaların daha kolay olması amacıyla, sürekli nicel verilerin frekans tabloları düzenlenmektedir. Açık uçlu gruplanmış verinin tanımlayıcı istatistiklerini hesaplamada bazı zorluklar ortaya çıkmaktadır. Ortalama ve standart sapmanın formüllerinin sınıf değerlerine, sınıf değerlerinin de sınıf aralıklarına dayanmasından dolayı, ilk sınıfın alt ve son sınıfın üst sınır değerini bilmek gerekir. Önceki çalışmalarda, bilinmeyen sınırları tahmin etmek için çeyrekler, ondalıklar, yüzdeler ve sınıf ara değerleri arası aralıklar kullanılmıştır. Bu çalışmada, açık uçlu gruplanmış verilerinde kitle ortalaması ve standart sapmasını tahmin etmek amacıyla dört yöntem önerilmiştir. Bu yöntemleri karşılaştırmak amacıyla kapsamlı bir Monte Carlo benzetimi uygulanmış ve sonuçlar CO2 emisyon verileri üzerinden tartışılmıştır.
Mean and Standard Deviation for Open-Ended Grouped Data
Frequency table of continuous quantitative data is arranged so that to describe the data better and convenience of numerical calculations. There are some difficulties to calculate the descriptive statistics of open-ended grouped data. Because the formulations of mean and standard deviation are based on midpoints, and midpoints are based on class intervals, it is necessary to know the lower-most and upper-most categories. In the previous studies, the interquartile, interdecile, interpercentile, and mid-distance ranges were used to estimate the unknown boundaries. This paper proposes four methods to estimate the population mean and standard deviation in open-ended group data. We conduct an extensive Monte Carlo simulation to compare these methods and the results are discussed over CO2 emission data.
___
- [1] Freund J.E., “Modern elementary statistics” 11th Ed., Pearson Education, Upper Saddle River, New Jersey, (2004).
- [2] Yilmaz A.E. and Saracbasi T., “The effect of changing scores for multi-way tables with open-ended ordered categories”, Hacettepe Journal of Mathematics and Statistics, 45(6): 1881-1890, (2016).
- [3] Nick T.G., “Descriptive Statistics”, Topics in biostatistics, Humana Press, Totowa, New Jersey, (2007).
- [4] Tukey J.W., “Exploratory data analysis”, Addison-Wesley, Massachusetts, (1977).
- [5] Frigge M., Hoaglin D.C. and Iglewicz, B., “Some implementations of boxplot”, The American Statisticians, 43(1): 50-54, (1989).
- [6] Yang J., Rahardja S. and Fränti P., “Outlier detection: How to threshold outlier scores?”, International Conference on Artificial Intelligence, Information Processing and Cloud Computing, 1-6, (2019).
- [7] Simmons J.P., Nelson L.D. and Simonsohn U., “False positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant”, Psychological Science, 22(11): 1359–1366, (2011).
- [8] Miller J., “Reaction time analysis with outlier exclusion: Bias varies with sample size”, The Quarterly Journal of Experimental Psychology, 43(4): 907– 912, (1991).
- [9] https://data.worldbank.org/indicator/EN.ATM.CO2E.P, The World Bank Data, “CO2 Emissions (metric tons per capita)”, (2019).