2015 Fiscal Year Research-status Report
Programming learner supports dependent on institutes
| Project/Area Number |
26730178
|
|---|
| Research Institution | Ritsumeikan University |
|---|
Principal Investigator |
デイン.テイ.ドン フウオン 立命館大学, 情報理工学部, 助手 (90712117)
|
|---|
| Project Period (FY) |
2014-04-01 – 2017-03-31
|
| Keywords | NMF / motivation / learning strategy / behavior / cosine similarity / matrix / vector / contextual inquiry |
|---|
| Outline of Annual Research Achievements |
To solve the personalized exercise in the distance learning, we aim to achieve the followings in this project;
(1) We establish data representation of portfolios expressing successful cases and learning behavior.
(2) We develop a method to calculate learning status.
(3) We compare the learning status with that of past students to provide personalized support.
In the proposal, we aimed to calculate the understanding level of each student. However, we have found what is actually important is not understanding level, but motivation and learning strategy to improve programming skills of students.
In 2015, we have developed a method to realize (2) and (3), because we attained (1) in 2014. For (2), we have achieved a calculation method using Non-negative matrix factorization(NMF). It figures out motivation and learning strategy of each student from his/her learning behavior. We confirmed the effectiveness of the method using data sampled from Japanese students. For (3), we have established a detection method of active students who can solve their difficulties by themselves. Once we can discriminate them, we can concentrate our efforts to supervise passive students. We have also applied the method to get motivation of novice farmers.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
First, we try to figure out the understanding level of each student. However, in the actual class, we have found many students get desperate when they feel hopeless. To detect these desperate students, we should identify motivation and learning strategy. We use NMF to figure out motivation and learning strategy from learning behavior of each student.
We have established a calculation method of motivation and learning strategy. We applied it to data sampled from Japanese students. We confirmed its effectiveness, comparing the calculation results with interview results for more than 40 students.
|
| Strategy for Future Research Activity |
When we confirm the effectiveness in Japan, we applied the method in the e-learning environment over the Internet. We believe we can apply the method in distance learning environments. However, it has revealed that the method strongly depends on data sampled from past students. To confirm effects of different educational institute, we are going to apply this method to an educational institute other than the Japanese one, we assume to evaluate the method in a Vietnamese institute.
In 2015, we have already tried the evaluation in Vietnam, but we failed because of problems in the bandwidth of the network between the two countries. In 2016, we will try one more time, after we solve the network problem.
|
| Causes of Carryover |
We planed to publish a transaction paper, but it took some more time to be accepted. Once it is accepted, we need registration fee for the paper. We planed to use the remaining budget for the registration fee.
|
| Expenditure Plan for Carryover Budget |
In 2016, we are going to publish the transaction paper. In addition, we plan to have another experiment in Japan and Vietnam. We are going to use the budget for the registration fee of the paper and the experiment.
|
