Reliability of Checking an Answer Given by a Mathematical Expression in Interactive Learning Systems

TL;DR

This paper proposes an enhanced answer checking method for interactive learning systems with mathematical notation, combining standard algorithms with pointwise checks, and analyzes its low error probability to improve reliability.

Contribution

It introduces a reinforced answer checking algorithm with an additional pointwise procedure and provides a probabilistic analysis of its error rate, enhancing reliability in mathematical answer verification.

Findings

Error probability is extremely low in common cases.

Additional checking improves answer verification reliability.

Method helps avoid sudden identity issues in answer checking.

Abstract

In this article we address the problem of automatic answer checking in interactive learning systems that support mathematical notation. This problem consists of the problem of establishing identities in formal mathematical systems and hence is formally unsolvable. However, there is a way to cope with the issue. We suggest to reinforce the standard algorithm for function comparison with an additional pointwise checking procedure. An error might appear in this case. The article provides a detailed analysis of the probability of this error. It appears that the error probability is extremely low in most common cases. Generally speaking, this means that such an additional checking procedure can be quite successfully used in order to support standard algorithms for functions comparison. The results, obtained in this article, help avoiding some sudden effects of the identity problem, and provide a way to estimate the reliability of answer checking procedure in interactive learning systems.