Difficulty and Typicality-Based Problem Recommendation Algorithm Implementation and Analysis

With the increasing popularity of "online learning", a method of finding effective learning materials from massive resources is urgently needed. By setting the "problems" in the online learning system as an example, this paper analyzes the "difficulty" of learning materials and proposes a computational method of difficulty. Then, based on the difficulty of problems and the concept of "typicality" in psychology, an algorithm suitable for problem recommendation is designed. The comparative experiments based on the actual data prove that this algorithm is superior to the existing methods both in the recommendation effectiveness and the time efficiency.