Item response prediction for incomplete response matrix using the EM-type item response theory with application to adaptive online ability evaluation system

Information Engineering Express

2018

Self Cite

One of the crucial issues in universities where a variety of enrolled students shall be educated to a level of university diploma policy is to identify students at risk for failing courses and/or dropping out early, to take care of them, and to reduce their risks. Using the recently developed follow-up program system aimed at helping students who need basic learning and aimed at assisting teachers who have to engage in teaching a variety of educational students, we can analyze the accumulated testing results in detail because the testings are performed every week to all the first-year undergraduate students. We have found that those who failed in the final examination show the much steeper decreasing trend of correct answer rates in the learning check testing comparing to those who were successful in the final examination. Although the subjects dealt with in this paper are limited to mathematics (calculus and linear algebra), this kind of system will easily be applied to other subjects such as STEM.

Section: Item Response Matrixmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Difference Between Successful and Failed Students Learned from Analytics of Weekly Learning Check Testing

Information Engineering Express

2018

Self Cite

“…To deal with the incomplete item response matrix, we propose to use the EM-type IRT [6]. In using Equation (2), it is common that the value of should be an indicator such that when successfully solved and otherwise.…”

Section: Matrix Completion By Using the Em-type Irtmentioning

confidence: 99%

“…Thus, we have to use matrix completion methods which can estimate the item difficulties and examinees' abilities altogether from incomplete item response matrices. The relevant studies are [5], [6], [7], [8], [9].…”

Section: Introductionmentioning

confidence: 99%

Automatically growing dually adaptive online IRT testing

2014 IEEE International Conference on Teaching, Assessment and Learning for Engineering (TALE)

Aizawa

2014

Self Cite

The item response theory (IRT) is widely used in many fields recently because it provides us the abilities of examinees and the difficulties of items simultaneously. Online IRT test systems are often equipped with the adaptive testing. Such a testing method selects the most appropriate items to examinees automatically. However, the item difficulty values remain the same even if the features of the examinees and the size of the items change. Calibration of the difficulty values is required in such cases, and the dually adaptive online IRT testing system works, where "dually adaptive" means that one is targeted to the adequate item selection and the other is targeted to the adjustment of the item difficulty values. Using this, the database in the system will automatically be growing and be updated. In this paper, we introduce such a novel system with applications.

“…Thus, in this paper, we use matrix completion methods which can estimate the item difficulties and examinees' abilities altogether from incomplete item response matrices. The relevant studies are [7], [8], [9], [10], [11].…”

Section: Introductionmentioning

confidence: 99%

A simulation study to the dually adaptive online IRT testing system

2014 IEEE International Conference on Teaching, Assessment and Learning for Engineering (TALE)

Tokusada

2014

Self Cite

The item response theory (IRT) gives us valuable information about the difficulties of items as well as the abilities of examinees. Adaptive testing using the IRT provides the most appropriate items to examinees automatically. However, the item parameters remain the same even if the features of the examinees and the number of examinees change. We propose to use the dually adaptive online IRT testing system, where "dually adaptive" means that one is targeted to the adequate item selection and the other is to the adjustment of the difficulty values for items. The key idea for this is to use the matrix completion method to the incomplete matrix. By using the simulation study, we found that the proposed method works well.