Randomization of Explainable and Interactive Simulated Annealing Strategy for Better Learning to Rank Models

Abstract This short paper is about the first Explainable and Interactive Learning to Rank (LTR) Package in Information Retrieval (IR). This application is based on Combining Simulated Annealing Strategy with (1+1)-Evolutionary Strategy (SAS-Rank) which was introduced before as learning algorithm for ranking in the previous studies. In this application, the ranking models for offspring and parent chromosomes were showed in the runing for each iteration. Furthermore, there are three options for changing the SAS-Rank parameters and seeing the evaluation results obtained for that. This application is the first application introduces interactive learning in the ranking problem domain for IR.