APPLICATION OF PREDICTIVE ANALYTIC FOR ITB STUDENT DROPOUT CLASSIFICATION USING DEEP LEARNING
Dropout is a problem of every college. One method that used to preven dropout is to build an early warning system. Deep learning is a technique to predict potential risk of dropout student with good accuracy efficiently. However, ITB still has not developed an early warning system for student with t...
Saved in:
| 主要作者: | |
|---|---|
| 格式: | Final Project |
| 語言: | Indonesia |
| 在線閱讀: | https://digilib.itb.ac.id/gdl/view/56552 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Institut Teknologi Bandung |
| 語言: | Indonesia |
| 總結: | Dropout is a problem of every college. One method that used to preven dropout is to build an early warning system. Deep learning is a technique to predict potential risk of dropout student with good accuracy efficiently. However, ITB still has not developed an early warning system for student with the risk of dropout using deep learning.
an
.
Several deep learning algorithms have been carried out in previous studies. Using various types of algorithm can provide different performance. In addition, features or attributes of data used to train model also have an effect. Therefore, it's necessary to
research types of algorithms and data that are suitable for ITB students to provide the best performance result.
In this final project, will try to build a model using Feed-Forward Neural Network and Recurrent Neural Network. The dataset used is data from undergraduate student of ITB. However, the data are limited only using student academic scores without demographic and background data. This data will be focus on grade point score of students
From the research result, best model is to use Feed-Forward Neural Network algorithm
by training student academic scores data and some its derivative attributes. Training data is supported by SMOTE method to overcome imbalanced data, resulting in model with average performance value precision and recall is 0.97 and 0.98, respectively. |
|---|