Abstract
Conference Title: 2018 IEEE Frontiers in Education Conference (FIE) Conference Start Date: 2018, Oct. 3 Conference End Date: 2018, Oct. 6 Conference Location: San Jose, CA, USA In this Innovative Practice, Work-in-Progress paper we present a new version of our web-based problem solving learning environment called CalcTutor. The key feature of the new CalcTutor system is that it organizes Mathematics problems based on the skills that they require and uses a machine learning technique (Performance Factors Analysis) to provide a difficulty estimate for each problem, personalized for each student in the system. This facility allows students to effectively engage in self- study by searching the library for problems that require specified skills. It ranks the problems found in the search by their difficulty for that particular student (the probability the student will solve them correctly on the first attempt). The current version of CalcTutor makes creating problems easy as well. It provides a simple and powerful question authoring tool which allows teachers to build both multiple choice questions and questions where the answers are mathematical functions of one variable.