Group recommender systems (GRS) identify items to recommend to a group by aggregating group members' individual preferences into a group profile. The preference aggregation strategy used to build the group profile can also be used for predicting the item that a group may decide to choose, i.e., by assuming that the group is applying exactly that strategy. However, predicting the choice of a group is challenging since the RS is not aware of the precise preference aggregation strategy that is going to be used by the group. Hence, the aim of this paper is to validate the research hypothesis that, by using a machine learning approach and a data set of observed group choices, it is possible to predict a group's final choice, better than by using a standard preference aggregation strategy. Inspired by Social Decision Scheme theory, which first tried to address the group choice prediction problem, we search for a group profile definition that, in conjunction with a machine learning model, can be used to accurately predict a group choice. Moreover, to cope with the data scarcity problem, we propose two data augmentation methods, which add synthetic group profiles to the training data, and we hypothesise they can further improve the choice prediction accuracy. We validate our research hypotheses by using a data set containing 282 participants organized in 79 groups. The experiments indicate that the proposed methods outperform baseline aggregation strategies when used for group choice prediction. The proposed method is robust with the presence of missing preference data and achieves a performance superior to what human can achieve on the group choice prediction task. Finally, the proposed data augmentation method can also improve the prediction accuracy. Our approach can be exploited in novel GRSs to identify the items that the group is likely to choose and help the group to make a better choice.

翻译：群体推荐系统（GRS）通过将群体成员的个体偏好聚合成群体画像，从而识别向群体推荐的物品。用于构建群体画像的偏好聚合策略也可用于预测群体可能决定选择的物品，即假设群体正在应用该特定策略。然而，预测群体选择具有挑战性，因为推荐系统不清楚群体将使用的精确偏好聚合策略。因此，本文旨在验证以下研究假设：通过采用机器学习方法和包含观察到的群体选择的数据集，可以比使用标准偏好聚合策略更好地预测群体的最终选择。受首次尝试解决群体选择预测问题的社会决策方案理论启发，我们寻找一种群体画像定义，结合机器学习模型，可用于准确预测群体选择。此外，为应对数据稀缺问题，我们提出了两种数据增强方法，向训练数据中添加合成群体画像，并假设它们能进一步提升选择预测精度。我们通过包含282名参与者（组成79个群体）的数据集验证研究假设。实验表明，在用于群体选择预测时，所提方法优于基线聚合策略。该方法在缺失偏好数据的情况下表现出鲁棒性，且在群体选择预测任务中实现了优于人类的表现。最后，所提数据增强方法也能提高预测精度。我们的方法可应用于新型GRS中，用于识别群体可能选择的物品，并帮助群体做出更优决策。