For an increasing number of forecasting models based on computational intelligence, one of the most prioritized requests refers to the model’s transparency, explainability and reproducibility. With the constant emergence of more complex investment instruments and strategies, challenges in financial time series forecasting are being amplified. Feature selection and aggregation are typical examples of such challenges. This paper examines the interpolative Boolean algebra (IBA) approach for discovery of optimal logical aggregation (LA) of key factors in Hedge Funds’ (HF) investment strategies. IBA polynomials that serve as logical aggregation functions, are obtained as a product of a structure vector (SV) and corresponding atomic elements. To obtain optimal aggregation function for analyzed time series, structure vectors are optimized by iterating over all combinations of elements. The proposed approach is applied to four major groups (factors) of candidate inputs, first separately and then jointly, on five distinctive HFs’ time series. The evaluation and robustness check are examined using standard multivariable linear regression and more complex, extreme gradient boosting algorithms. Lastly, feature aggregation using optimal IBA logical functions are benchmarked against original, non-restricted inputs. Test error analysis has demonstrated that IBA-based feature aggregation reduces errors, for most of the analyzed time series, when compared to the original feature set.