Summary
The RE-WISE method is based on the mathematic model of forgetting and learning.
Forgetting models the formula for Retrievability – ability to recollect a Fact:
RET = 1/2^{m*SN/STAB}
- STAB - Stability is a number which characterizes how long a student will remember the Fact.
- m is the number of days from the last repetition of the Fact.
- SN is a constant implemented so the Stability equals the number of days recommended before the next repetition.
Learning models the formula for Stability growth:
STAB_{0} = iniStab ~= 1
STAB_{i+1} = STAB_{i} * CHNG(RET)
- IniStab is the Initial Stability at the beginning of learning. Usually, it equals 1.
- STAB_{i} is the Stability in repetition in the i. order
- CHNG is a function depending on RET. It is optimal to repeat at the moment when RET drops by approx. 30%. At this point, the CHNG function is at maximum. This maximum is called the Base Factor and it usually equals to 2.
- IniStab and Base Factor are adapted parameters. RE-WISE adjusts their values so the expected values of Retrievability correspond as much as possible to the achieved point evaluation.
Under “normal circumstances”, thus:
- IniStab = 1
- Base Factor = 2
- The student reviews in the recommended days and always gets an evaluation of 3 points on the scale 0-4
... the Stability grows in geometric progression and the recommended intervals of repetition are: 1, 2, 4, 8, 16, 32... days.
The implemented model adapts behavior of an algorithm to various difficulties of Facts and the various memories of students. It can also handle if the student does not review at the recommended intervals.