A retrospective classification model of historical success patterns in Hollywood actors using multi-criteria decision-making and machine learning

Iranian Journal of Science and Technology, Transactions of Electrical Engineering · Springer Nature · 2026-11-18 · DOI: 10.1007/s44163-026-02782-z

Аннотация

Actor success has been studied mostly after the fact and rarely at the level of whole careers. This study reconstructs a census of 2618 Hollywood performers from Wikidata and the official IMDb datasets, covering 1118 performers with at least one recorded Academy Award, Golden Globe, or BAFTA nomination or win and 1500 highly credited performers without such records. Six award indicators and a capped employment rate were weighted with SWARA on the judgments of 18 cinema-industry experts and aggregated into a historical success composite (HSC), cut at 0.25, 0.50, and 0.75 into four ordinal levels. Eight classifiers, including two dummy baselines, were compared under repeated stratified 5-fold cross-validation on 1544 complete records with accuracy, balanced accuracy, macro and weighted F1 and the Matthews correlation coefficient (MCC). Tuned support vector machines, random forests, and multinomial logistic regression reached accuracies of 0.76 and MCC values near 0.59 against a majority baseline of 0.537, while balanced accuracies of 0.44 to 0.51 and out-of-fold recalls below 0.17 for the two upper levels showed that these tiers are not separable from award-free career attributes. The analysis is retrospective and descriptive and makes no forecasting claim for careers without established award records.

Читать у издателя →

Похожие статьи