2018Unpublished venueRequires access

Penerapan Distribusi Random-Clumped Binomial Untuk Mengatasi Masalah Overdispersi Dalam Data Biner

Verdina Ambarwati, Nusar Hajarisman

Open publisher page 0 citations

Abstract

Abstract . Binary data is the generated data from the numerical mapping of an event that has only two possible events. Where the event is considered a success is symbolized by the number 1 whereas the event fail is symbolized by the number 0. The number of success events is followed by the binomial distribution if it meets the assumption that response variables are independent and have equal success chances. In practice, there are often violations of assumptions. This is an indication of an overdispersion problem that is the condition in which the variance of the T response (the number of successful events) exceeds the variance of the Binomial. Data can be expressed as overdispersion when the ratio between devides or chi squares of Pearson to its degree of freedom is more than 1. The Random-Clumped Binomial distribution discussed in this thesis is one method that can be used to overcome the overdispersion problem in the data following the binomial distribution. Random-Clumped Binomial Distribution is then applied to find out whether the distribution can overcome the problem of overdispersion in the data of Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy at Courses in Semester 1 and Semester 2. Results obtained that the distribution of Random-Clumped Binomial can overcome the problem of overdispersion that occurs in the data, and obtained the opportunity that the possibilty of  Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy get grades not graduated in the semester 1 and 2 courses is 0.1016 with the correlation between the value of one course with the value of the eye other lectures in semesters 1 and 2 of Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy amounted 0.5661. Keywords: Binary Data, Binomial Distribution, Random-Clumped Binomial Distribution, Maximum Likelihood Method, Devians and Chi Squares Pearson. Abstrak .Data biner merupakan data yang dibangkitkan dari pemetaan numerik dari suatu kejadian yang hanya memiliki dua kemungkinan kejadian. Dimana kejadian yang dianggap sebagai kejadian “sukses” disimbolkan dengan angka 1 sedangkan kejadian “gagal” disimbolkan dengan angka 0. Banyaknya peristiwa “sukses” dinyatakan mengikuti distribusi binomial apabila memenuhi asumsi bahwa variabel responnya saling bebas dan mempunyai peluang “sukses” yang sama. Pada praktiknya sering terjadi pelanggaran asumsi. Hal tersebut merupakan salah satu indikasi adanya masalah overdispersi yaitu kondisi dimana varians dari respon T (banyaknya kejadian sukses) melebihi varians dari Binomial. Data dapat dinyatakan mengalami masalah overdispersi ketika rasio antara nilai devians atau chi kuadrat Pearson terhadap derajat bebasnya lebih dari 1. Distribusi Random-Clumped Binomial yang dibahas dalam skripsi ini  merupakan salah satu metode yang dapat digunakan untuk mengatasi masalah overdispersi pada data yang mengikuti distribusi binomial. Distribusi Random-Clumped Binomial kemudian diaplikasikan untuk mengetahui apakah distribusi tersebut dapat mengatasi masalah overdispersi pada data Status Kelulusan Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung pada Mata Kuliah di Semester 1 dan Semester 2. Hasil yang diperoleh bahwa distribusi Random-Clumped Binomial dapat mengatasi masalah overdispersi yang terjadi pada data, dan diperoleh peluang bahwa Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung memperoleh nilai tidak lulus pada mata kuliah semester 1 dan 2 adalah 0.1016 dengan korelasi antara nilai mata kuliah yang satu dengan nilai mata kuliah yang lainnya pada semester 1 dan 2 dari Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung sebesar 0.5661. Kata kunci : Data biner, Distribusi Binomial, Distribusi Random-Clumped Binomial, Metode Kemungkinan Maksimum, Devians  dan Chi Kuadrat Pearson.

About this research paper

What this paper is about

Abstract . Binary data is the generated data from the numerical mapping of an event that has only two possible events. Where the event is considered a success is symbolized by the number 1 whereas the event fail is symbolized by the number 0. The number of success events is followed by the binomial distribution if it meets the assumption that response variables are independent and have equal success chances. In practice, there are often violations of assumptions. This is an indication of an overdispersion problem that is the condition in which the variance of the T response (the number of successful events) exceeds the variance of the Binomial. Data can be expressed as overdispersion when the ratio between devides or chi squares of Pearson to its degree of freedom is more than 1. The Random-Clumped Binomial distribution discussed in this thesis is one method that can be used to overcome the overdispersion problem in the data following the binomial distribution. Random-Clumped Binomial Distribution is then applied to find out whether the distribution can overcome the problem of overdispersion in the data of Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy at Courses in Semester 1 and Semester 2. Results obtained that the distribution of Random-Clumped Binomial can overcome the problem of overdispersion that occurs in the data, and obtained the opportunity that the possibilty of  Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy get grades not graduated in the semester 1 and 2 courses is 0.1016 with the correlation between the value of one course with the value of the eye other lectures in semesters 1 and 2 of Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy amounted 0.5661. Keywords: Binary Data, Binomial Distribution, Random-Clumped Binomial Distribution, Maximum Likelihood Method, Devians and Chi Squares Pearson. Abstrak .Data biner merupakan data yang dibangkitkan dari pemetaan numerik dari suatu kejadian yang hanya memiliki dua kemungkinan kejadian. Dimana kejadian yang dianggap sebagai kejadian “sukses” disimbolkan dengan angka 1 sedangkan kejadian “gagal” disimbolkan dengan angka 0. Banyaknya peristiwa “sukses” dinyatakan mengikuti distribusi binomial apabila memenuhi asumsi bahwa variabel responnya saling bebas dan mempunyai peluang “sukses” yang sama. Pada praktiknya sering terjadi pelanggaran asumsi. Hal tersebut merupakan salah satu indikasi adanya masalah overdispersi yaitu kondisi dimana varians dari respon T (banyaknya kejadian sukses) melebihi varians dari Binomial. Data dapat dinyatakan mengalami masalah overdispersi ketika rasio antara nilai devians atau chi kuadrat Pearson terhadap derajat bebasnya lebih dari 1. Distribusi Random-Clumped Binomial yang dibahas dalam skripsi ini  merupakan salah satu metode yang dapat digunakan untuk mengatasi masalah overdispersi pada data yang mengikuti distribusi binomial. Distribusi Random-Clumped Binomial kemudian diaplikasikan untuk mengetahui apakah distribusi tersebut dapat mengatasi masalah overdispersi pada data Status Kelulusan Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung pada Mata Kuliah di Semester 1 dan Semester 2. Hasil yang diperoleh bahwa distribusi Random-Clumped Binomial dapat mengatasi masalah overdispersi yang terjadi pada data, dan diperoleh peluang bahwa Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung memperoleh nilai tidak lulus pada mata kuliah semester 1 dan 2 adalah 0.1016 dengan korelasi antara nilai mata kuliah yang satu dengan nilai mata kuliah yang lainnya pada semester 1 dan 2 dari Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung sebesar 0.5661. Kata kunci : Data biner, Distribusi Binomial, Distribusi Random-Clumped Binomial, Metode Kemungkinan Maksimum, Devians  dan Chi Kuadrat Pearson.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Abstract . Binary data is the generated data from the numerical mapping of an event that has only two possible events. Where the event is considered a success is symbolized by the number 1 whereas the event fail is symbolized by the number 0. The number of success events is followed by the binomial distribution if it meets the assumption that response variables are independent and have equal success chances. In practice, there are often violations of assumptions. This is an indication of an overdispersion problem that is the condition in which the variance of the T response (the number of successful events) exceeds the variance of the Binomial. Data can be expressed as overdispersion when the ratio between devides or chi squares of Pearson to its degree of freedom is more than 1. The Random-Clumped Binomial distribution discussed in this thesis is one method that can be used to overcome the overdispersion problem in the data following the binomial distribution. Random-Clumped Binomial Distribution is then applied to find out whether the distribution can overcome the problem of overdispersion in the data of Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy at Courses in Semester 1 and Semester 2. Results obtained that the distribution of Random-Clumped Binomial can overcome the problem of overdispersion that occurs in the data, and obtained the opportunity that the possibilty of  Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy get grades not graduated in the semester 1 and 2 courses is 0.1016 with the correlation between the value of one course with the value of the eye other lectures in semesters 1 and 2 of Student Pass Status Year 2016/2017 in Statistics Majority, Science and Mathematics Faculty, Bandung Islamic Univeristy amounted 0.5661. Keywords: Binary Data, Binomial Distribution, Random-Clumped Binomial Distribution, Maximum Likelihood Method, Devians and Chi Squares Pearson. Abstrak .Data biner merupakan data yang dibangkitkan dari pemetaan numerik dari suatu kejadian yang hanya memiliki dua kemungkinan kejadian. Dimana kejadian yang dianggap sebagai kejadian “sukses” disimbolkan dengan angka 1 sedangkan kejadian “gagal” disimbolkan dengan angka 0. Banyaknya peristiwa “sukses” dinyatakan mengikuti distribusi binomial apabila memenuhi asumsi bahwa variabel responnya saling bebas dan mempunyai peluang “sukses” yang sama. Pada praktiknya sering terjadi pelanggaran asumsi. Hal tersebut merupakan salah satu indikasi adanya masalah overdispersi yaitu kondisi dimana varians dari respon T (banyaknya kejadian sukses) melebihi varians dari Binomial. Data dapat dinyatakan mengalami masalah overdispersi ketika rasio antara nilai devians atau chi kuadrat Pearson terhadap derajat bebasnya lebih dari 1. Distribusi Random-Clumped Binomial yang dibahas dalam skripsi ini  merupakan salah satu metode yang dapat digunakan untuk mengatasi masalah overdispersi pada data yang mengikuti distribusi binomial. Distribusi Random-Clumped Binomial kemudian diaplikasikan untuk mengetahui apakah distribusi tersebut dapat mengatasi masalah overdispersi pada data Status Kelulusan Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung pada Mata Kuliah di Semester 1 dan Semester 2. Hasil yang diperoleh bahwa distribusi Random-Clumped Binomial dapat mengatasi masalah overdispersi yang terjadi pada data, dan diperoleh peluang bahwa Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung memperoleh nilai tidak lulus pada mata kuliah semester 1 dan 2 adalah 0.1016 dengan korelasi antara nilai mata kuliah yang satu dengan nilai mata kuliah yang lainnya pada semester 1 dan 2 dari Mahasiswa Tahun 2016/2017 Program Studi Statistika FMIPA Universitas Islam Bandung sebesar 0.5661. Kata kunci : Data biner, Distribusi Binomial, Distribusi Random-Clumped Binomial, Metode Kemungkinan Maksimum, Devians  dan Chi Kuadrat Pearson.

Key concepts: Overdispersion, Mathematics, Quasi-likelihood, Statistics, Negative multinomial distribution, Negative binomial distribution, Beta-binomial distribution, Binomial distribution

Back to paper searchBrowse research topicsOriginal source
Penerapan Distribusi Random-Clumped Binomial Untuk Mengatasi Masalah Overdispersi Dalam Data Biner — Research Paper | ScholarLens