• Hakeem Opeyemi Siyanbola

Online consumer ratings are widely used to summarize product evaluations, but reported prediction errors depend on the unit of analysis. Using 1,048,340 BeerAdvocate reviews covering 42,716 beers and 28,761 reviewers, this study compares reconstruction of contemporaneously recorded overall ratings from sensory scores under two representations: beer-level averages and individual reviews. Linear Regression, Random Forest, and a validation-weighted ensemble were evaluated using beer-disjoint development and test partitions. The ensemble achieved MAE = 0.194, RMSE = 0.280, and R² = 0.790 for beer-level aggregates, compared with MAE = 0.305, RMSE = 0.405, and R² = 0.682 for individual reviews. Across matched model classes, raw-review RMSE was 38.8–45.6% higher. These errors are not like-for-like estimates for an identical prediction target: aggregation changes outcomes, sensory predictors, observation counts, and weighting (one beer versus one review per observation). Permutation importance identified taste, palate, and aroma as the leading predictors. A crossed random-intercept model attributed 28.9% of rating variance to beer identity, 7.0% to reviewer identity, and 64.1% to residual review-level variability. The findings show that product-level consensus is easier to reconstruct in this setting, not that personalized recommendations are more accurate. Retailers and review platforms should evaluate consensus estimation and individual-response reconstruction separately and display rating dispersion alongside product averages.

Online consumer reviews; consumer preference; rating aggregation; consumer heterogeneity; recommender systems; retail analytics.

"From Consumer Consensus to Individual Preference: How Rating Aggregation Shapes Predictive Accuracy in Online ReviewsMER REVIEWS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 10, page no.b643-b658, October-2026, Available :http://www.jetir.org/papers/JETIR2610164.pdf

"From Consumer Consensus to Individual Preference: How Rating Aggregation Shapes Predictive Accuracy in Online ReviewsMER REVIEWS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 10, page no. ppb643-b658, October-2026, Available at : http://www.jetir.org/papers/JETIR2610164.pdf