Here we present a new open source implementation of routines to perform fitting data to logistic functions based on Item Response Theory using scipy.optimize. The field of Item Response Theory (IRT) is concerned with accurately measuring aptitude based on responses to itemized evaluations (often a set of multiple-choice questions). At Pluralsight, LLC (and previously Smarterer, Inc) we have developed a proprietary scoring algorithm that vastly improves the speed at which we can evaluate these "skills" very accurately. However, because it is useful to compare our system with existing standards in testing, we have developed our own implementation of the IRT maximum likelihood estimation routines to fit 1, 2, or 3 parameter item characteristic curves (ICC) which are logistic functions. We have recently shared these in an Apache 2.0 licensed Python package (irt_parameter_estimation).

---