Mustafa Kellekci

Mr. Mustafa Kellekci

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
15 September 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MKMustafa Kellekci
Mustafa Kellekci
UCUfuk Cebeci
Ufuk Cebeci
Onur DoganOnur Dogan
Onur Dogan
Early-stage start-up selection is a critical yet challenging task for venture capital (VC) investors due to high uncertainty, limited historical data, and rapidly evolving business environments. Traditional evaluation processes often fall short in systematically handling multiple qualitative and uncertain factors that influence start-up success. As a result, there is a growing demand for robust decision models that can support VC firms in identifying promising early-stage ventures more accurately and consistently. This study presents a hybrid fuzzy multi-criteria decision-making approach tailored to the needs of venture capital investment under uncertainty. The model integrates expert judgment using the proportional spherical fuzzy AHP method to evaluate the relative importance of key dimensions. Then, spherical fuzzy TOPSIS is applied to rank investment alternatives based on their overall performance rankings. The proposed framework enables VC decision-makers to incorporate both subjective insights and data ambiguity in a structured and transparent way. It offers a practical tool to enhance the reliability of early-stage investment evaluations and improve the effectiveness of venture capital portfolio strategies.

21 November 2025
Verified
Book Chapter
Closed access
MKMustafa Kellekci
Mustafa Kellekci
UCUfuk Cebeci
Ufuk Cebeci
Onur DoganOnur Dogan
Onur Dogan
This chapter proposes a practical framework for venture capital (VC) investment decisions that combines multi-criteria decision making (MCDM) with machine learning (ML) in a human-in-the-loop setting. It focuses on early-stage VC, where decisions must be made under high uncertainty, limited data, and changing definitions of startup success. MCDM is used to capture expert value judgments as explicit criteria and weights, while ML models estimate criterion scores from observable signals such as innovation, team, traction, governance, and networks. The chapter explains the two-way integration between MCDM and ML, shows how explainable AI artifacts support auditability and learning, and illustrates, through case studies, how hybrid rankings can be turned into personalized, constraint-aware recommendation engines for different investor profiles.

23 March 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MKMustafa Kellekci
Mustafa Kellekci
UCUfuk Cebeci
Ufuk Cebeci
Onur DoganOnur Dogan
Onur Dogan
Many real-world screening tasks in venture capital must rank large start-up candidate pools under conditions of tight review capacity, time-varying information, and rare investment success outcomes. When datasets are constructed retrospectively, post-decision updates can leak into features and inflate performance, especially with random splits. This study proposes a leakage-aware, time-based evaluation framework for capacity-constrained screening formulated as a top-K ranking problem. Using a dataset of 117,141 early-stage firms as an empirical testbed, features were constructed strictly as of a reference time t0, a 180-day temporal embargo was enforced around the train–test boundary, and generalization was assessed with time-ordered splits. Because venture capital decisions are made on a shortlist, evaluation emphasizes ranking quality using PR-AUC, Lift@K, Precision@K/Recall@K, and NDCG@K, reported with bootstrap confidence intervals. Under this leakage-aware protocol and with strong class imbalance, maturity-related signals achieve the strongest PR-AUC (0.0144), while team and combined signals yield the best top-50 shortlist concentration. Finally, probability calibration substantially improves reliability for threshold planning (Brier score reduced from 0.0972 to 0.0161 with sigmoid calibration) while leaving ranking essentially unchanged. Overall, the study provides a leakage-aware evaluation template and an interpretable baseline for time-dependent venture capital screening tasks involving start-up selection, investment success prediction, leakage risk, and limited review capacity.