Peter Korošec

Journal articlesshow all

Beyond Landscape Analysis: DynamoRep Features For Capturing Algorithm-Problem Interaction In Single-Objective Continuous Optimization Evolutionary Computation, ECJ (2025)
Using Machine Learning Methods to Assess Module Performance Contribution in Modular Optimization Frameworks, ECJ (2024)
Opt2Vec - A continuous optimization problem representation based on the algorithm: a case study on problem classification, INFORM SCIENCES (2024)

Conference papersshow all

Adaptive Estimation of the Number of Algorithm Runs in Stochastic Optimization, GECCO 2025
Per-Run Algorithm Performance Improvement Forecasting Using Exploratory Landscape Analysis Features: A Case Study in Single-Objective Black-Box Optimization, GECCO 2024

Books & Chaptersshow all

Reliability Improvements for In-Wheel Motor, Recent Advances in Microelectronics Reliability, Springer (2024)
Deep Statistical Comparison for Meta-heuristic Stochastic Optimization Algorithms, Deep Statistical Comparison for Meta-heuristic Stochastic Optimization Algorithms, Springer Cham (2022)

Other publicationsshow recent

How to Perform a Proper Statistical Study Analysis? Where we Started and Where are we Now in Statistical Performance Assessment Approaches for Stochastic Optimization Algorithms? (2018)
Stigmergy as an approach to metaheuristic optimization (2006)
Metaheuristic solving of the optimization problem with ant colonies (2004)
Evaluation of heuristic partitioning algorithms (2001)

Less is more: Selecting the right benchmarking set of data for time series classification

T. Eftimov, G. Petelin, G. Cenikj, A. Kostovska, G. Ispirova, P. Korоšec, J. Bogatinovski
DOI: https://doi.org/10.1016/j.eswa.2022.116871

abstract:
In this paper, we have proposed a new pipeline for landscape analysis of time-series machine learning datasets that enables us to better understand a benchmarking problem landscape, allows us to select a diverse benchmark datasets portfolio, and reduces the presence of performance assessment bias via bootstrapping evaluation. Combining a large multi-domain representation corpus of time-series specific features and the results of a large empirical study of time-series classification (TSC) benchmark, we showcase the capability of the pipeline to point out issues with non-redundancy and representativeness in the benchmark. By observing discrepancy between the empirical results of the bootstrap evaluation and recently adopted practices in TSC literature when introducing novel methods, we warn on the potentially harmful effects of tuning the methods on certain parts of the landscape (unless this is an explicit and desired goal of the study). Finally, we propose a set of datasets uniformly distributed across the landscape space one should consider when benchmarking novel TSC methods.