Determining the Effect of Curriculum and Facilities on Academic Achievement Using Data Mining Approach
Education domain offers many interest and challenge in data mining applications that potentially identified as a tool to help both educators and students to improve the quality of education system. Data Mining applies modern statistical and computational technologies to the problem of finding usefu...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2008
|
| Subjects: | |
| Online Access: | http://etd.uum.edu.my/895/ http://etd.uum.edu.my/895/1/Surenah_Sedek.pdf http://etd.uum.edu.my/895/2/Surenah_Sedek.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Education domain offers many interest and challenge in data mining applications that potentially identified as a tool to help both educators and students to improve the quality of education system. Data Mining applies modern statistical and computational technologies to the problem of finding
useful hidden patterns within large databases. Thus, this study applied data mining technique to identify the hidden information that affects the academic achievement among respondents. The respondents for this study are consists of
all public university students which complete their study within year 2007. The questionnaire that has been used in this study was adopted from Kajian Pengesanan Graduan, Kementerian Pengajian Tinggi and it focuses on curriculum and facilities that have been provided by universities. The aims of this study is to determine whether the curriculum and facilities that provided by university has an effect on students academic achievement. 55,315 respondents data were used for descriptive task while 39,801 data for predictive task. Both data mining approaches, namely the descriptive and predictive have been utilized to perform the analysis prior to build the model. For descriptive purposes, frequency, cross tabulation and correlation coefficients were computed to check whether significant correlation exists. For predictive modeling, logistic regression and neural network were used. Statistical Pakages for Sosial Science (SPSS) was used for regression technique and Statistical Analytical Software (SAS) for Neural Network modeling. Then, the online questionnaire was integrated with Neural Network model to predict future student academic achievement. The findings in this study suggest neural network is the best model compared to logistic regression to measure the effect of curriculum and facilities on academic achievement. The highest accuracy from neural network is 89.47%, when demographics and curriculum become
the contributing variables to academic achievement. Most of the neural network model accuracy is over than 80% while logistic regression accuracy is below than 50 %. |
|---|