Identification of the Cause of Errors in Solving Mathematical Problems
Mutants, Distractors, Mathematics, STI, Errors, Difficulties, Step-by-step
Learning mathematics can be a significant challenge for students around the world. The difficulties encountered by these students vary, such as lack of attention, methodological problems, lack of mastery of prerequisite knowledge, reading difficulties, personal issues, among others. Despite being complex, these difficulties often manifest themselves in specific errors in solving mathematical problems, allowing experts to identify them and associate them with probable causes. In this context, common errors stand out, such as operator substitutions, rounding errors, incorrect results of operations, among others. These errors can be mapped and generalized since they are integral parts of students' solutions. Thus, intelligent systems, such as Intelligent Tutoring Systems (ITS), can be developed to address these difficulties by identifying errors and providing feedback to teachers and students themselves. Based on the above, this work aims to propose a model for generalizing common errors to be applied in the step-by-step identification of error origins. To achieve this, the model will utilize the concept of mutants to generate distractors that will serve as parameters for identifying the source of the problems. To gather relevant data for this study, some research efforts have focused on assessing the state of the art of ITS applied to mathematics in the Brazilian and international scenarios. Furthermore, exploratory studies have been conducted to identify common errors that can be mapped for the generation of the mutation model. Next, the presentation of the mutant modeling is carried out, along with the description of the architecture of the ITS for mathematics and studies that seek to validate it. The main research hypotheses indicate that the use of mutant modeling applied to mathematics through an ITS allows for greater dynamism in creating error scenarios, and they can be associated with problems that go beyond the analysis of the test. Another hypothesis is that feedback based on distractors generated by the mutation model, combined with step-by-step analysis of students' responses, allows for more detailed identification of the error location, facilitating the generation of feedback from the ITS.