Bu kaydın yasal hükümlere uygun olmadığını düşünüyorsanız lütfen sayfa sonundaki Hata Bildir bağlantısını takip ederek bildirimde bulununuz. Kayıtlar ilgili üniversite yöneticileri tarafından eklenmektedir. Nadiren de olsa kayıtlarla ilgili hatalar oluşabilmektedir. MİTOS internet üzerindeki herhangi bir ödev sitesi değildir!

Predicting Student Success: A Case Study Based On The Transcript And Personal Data Of The Graduated Students At Computer Engıneering Department, Atılım Unıversity, Turkey

Oluşturulma Tarihi: 21-03-2019

Niteleme Bilgileri

Tür: Tez

Alt Tür: Yüksek Lisans Tezi

Yayınlanma Durumu: Yayınlanmamış

Dosya Biçimi: PDF

Dil: İngilizce

Konu(lar): TEKNOLOJİ,

Yazar(lar): Ceyhan, Ulaş Ozan (Yazar),

Emeği Geçen(ler): Karakaya, Murat (Danışman),

Anahtar Kelimeler

machine learning, educational data mining, data mining, prediction, learninganalytic.


Özet

In recent years, Educational Data Mining (EDM) has become more popular in dataanalysis projects. Institutes try to improve their educational quality as well as to investin analyzing educational data. Predicting student grades is a significant challenge inEDM, and also it has lots of benefits for improving quality in the education process. Inthis study, we aim at predicting the student success in the selected courses consideringtheir transcript and personal data. For this reason, we applied various Machine Learning(ML) algorithms on the graduated student data. We developed several conceptsfor analyzing students success at the selected courses. First, we define three metrics tomeasure the student success. These student success metrics are grade letters, successgroups and fail-pass state. Furthermore, we created four dierent data sets from thegraduated student data as inputs to the selected ML algorithms. Results of this studyindicate that high or low student grade letters can be predicted better when comparedto the moderate grade letters. Similarly, according to the success group metric; lowerand higher success groups of the students can be predicted with higher accuracy comparedto the average success group. For the last success metric, the prediction resultsare far better for the passed students than predicting the failed students. Consideringthe four input data sets, we could not locate considerable dierences in predictionivsuccess. However, the data set created by the student personal data generates lowerprediction success compared the rest of the data sets. The prediction success for theselected courses is observed to be increasing for the courses at the last two semestersof the curriculum. The details of the findings and their possible causes are analyzedand discussed in the related chapters. We believe that the results of this study canserve as a foundation to build a student recommendation system to predict their futurecourse success.


İçindekiler



Açıklamalar



Haklar



Notlar



Kaynakça


Atıf Yapanlar

Gözat Sayfasına Dön

 

Sosyal Medya ve Araçlar

İstatistikler

  • Kayıt
    • Bu ay: 2
    • Toplam: 2437
  • Online
    • Ziyaretçi: 9
    • Üye: 0
    • Toplam: 9

Detaylı İstatistikler