"Give me an easier related problem."
⇧
⇖
⇘
Learning mathematics is generally thought to be difficult. It is especially difficult if you don't know the prerequisites: just trying a to solve a problem isn't going to work without access to source materials and some ideas about how the components fit together. Can we use text and hypertext analysis to provide automatic connections of textbook problems to the relevant readings?
- Research Problem
- Document any learning impact from showing prerequisites and other connections in detail.
- "If there is a problem you can't solve, look for an
easier related problem that you also can't solve" is one
way to paraphrase Polya's main suggestion from his famous
book "How to Solve it".
- Implementation
- The thought is to use the Concept Forest algorithm to detect related texts, augmenting WordNet with synonyms etc. from PlanetMath's thesaurus. The algorithm should be adjusted to work with hypertextual relationships (automatically detected links, explicit connections suggested by users), and hopefully should learn from corrections when suggestions are poor.
- The approach uses fairly standard Linked Data representations. We MAY make use of MathML/OMDoc representations (or suggest that for future work).
- Analytical techniques
- We hope to document a learning increase (using the methodology from earlier paper, versus the Web 2.0 version of the results).
- Future work and implications
- Being able to solve a problem isn't the same as learning how to solve a problem. Suggesting related material that helps people solve problems might not be the "best" way to help them learn. Accordingly the follow-up paper (next section) is about making recommendations not about related content, but about relevant use of content.
- References