Skip to content

Latest commit

 

History

History
69 lines (49 loc) · 3.42 KB

README.md

File metadata and controls

69 lines (49 loc) · 3.42 KB

Math Problem Dataset

This is a geometry math problem dataset that focus on the quantitaive reasoning (stand upon the Descartes's mind to measure geometry). It can be used for different research fields: student modeling, natural language processing, scaffolding synthesis.

Three dimensions of data:

  1. Math problems: Problems.json

  2. Math concepts: Topics.json + Concepts.json

  3. Math scaffolding per problem: Scaffoldings.json (preliminary structure)


Dimension 1:

Format of Problems.json:

Column Name Column Type Column Explanation
id string GUID
problem string Text Problem
semantic array Annotated Math Semantic of natural language math problem
concept array Annotated Concepts covered in the problem (Concepts.json)
topic string Annotated Math Problem Topic (Topics.json)
solution array Annotate problem solution

Dimension 2:

Format of Topics.json:

Column Name Column Type Column Explanation
topic string Topic Name
dependencies array Dependent Topics

Format of Concepts.json:

Column Name Column Type Column Explanation
concept string Concept Name
topic string Topic Name
dependencies array Dependent Concepts

Dimension 3:

Format of Scaffolds.json:

Column Name Column Type Column Explanation
problem-id string corresponds to ID in problems.json
problem-text string question being asked
steps list of records step in scaffold (see below)

"step" record:

Column Name Column Type Column Explanation
id string unique identifier
type string summary,meta,concept-scaffold, procedural-scaffold,solution
text string embed text of scaffold step
dependencies list of list map all potential dependency step
concepts-used list of strings embed all topics employed by step
procedurals list of steps recursively add steps into the current step