Ensuring fairness and transparency in academic recruitment remains challenging, as traditional peer review faces limitations due to potential bias and subjectivity. Addressing this issue, Zeynep Karaman, the first graduate student of Bogazici University’s the Data Science and Artificial Intelligence (DSAI) Institute, has developed an AI-based approach in her master’s thesis to measure academic quality to support academic evaluation process. The study was supervised by Dr. Ercan Atam and Dr. Şaziye Betül Özateş.

This research introduces the Academic Quality Index (AQI) - a metric that combines established bibliometric indicators with three novel features: citations per paper per person (cpppp), maximum citations as first author and average author position (aap). Using a dataset of Nobel Prize laureates in Physics, Chemistry, Physiology or Medicine, and Economics, alongside full professors from mid-ranked universities, two machine learning models - Siamese Neural Networks and XGBoost - were trained on 2,331 researchers (with more than 296,000 publications). Both models achieved around 90% accuracy, with XGBoost slightly outperforming the neural network. Significantly, two of the features, cpppp and aap, emerged as top predictors of academic quality.

The findings pave the way for the further integration of AI into academic evaluation systems, enabling faster and more data-driven decision-making. By combining AQI scores with traditional peer review, this research has the potential to enhance academic evaluation processes.