Modeling Collaborative Effects in Bollywood Film Profitability Using Synergy Index

Abstract

This study introduces a novel feature engineering framework for analysing the influence of collaborative relationships and economic factors on film profitability in the Bollywood film industry. Traditional models often treat actor and director effects independently, ignoring the relational impact of specific collaborations, and frequently fail to account for the temporal depreciation of currency value in longitudinal datasets. To address these gaps, we propose Synergy Index, a metric that quantifies the joint performance of each actor–director pair relative to their individual historical averages. Furthermore, to ensure financial consistency across decades, we implement an inflation-adjustment mechanism that normalizes production budgets to a common base year using historical CPI data. The model additionally incorporates robust handling of multi-actor films and utilizes Bayesian-style shrinkage to avoid overfitting in sparse cases. We evaluate this framework using five state-of-the-art regression algorithms (Random Forest, XGBoost, LightGBM, CatBoost, and HistGradientBoosting). Feature importance analysis demonstrates that the Synergy Index consistently emerges as a significant predictor of financial success alongside normalized budget constraints. The methodology provides new insights into collaborative dynamics in Bollywood and forms a foundation for advanced, economically adjusted relational modelling in film analytics.