出于对协同效应的追求，近年来企业逐渐将团队作为其组织架构中最基本的组织形式。为了提升企业绩效表现，在大数据时代背景下，利用数据分析手段打造高效团队、实现协同效应、最大化团队绩效成为了重要的研究课题。然而，由于协同效应测量较为困难，现有研究框架中往往主要考虑团队成员或候选成员的个体信息，团队成员的协同效应更多地作为研究结果的解释因素出现，而未将其纳入整体研究框架中。而在现有对协同效应的测量或估计的研究中，现有文献主要依据团队候选成员的个体性格特质，并基于相应性格特质的组合对团队的协同效应进行测量或估计，从团队成员之间的合作历史角度出发对其进行测量或估计的研究相对较少。本文从数据可得性、研究现状契合度和产业背景出发，选择NBA（美国职业篮球联赛）作为研究情境，利用数据分析手段，从团队成员之间的合作历史角度出发量化团队的协同效应，研究考虑协同效应的团队绩效预测方法，从而指导团队构建。具体而言，本文提出了篮球阵容协同效应的测度，量化定义了篮球阵容团队协同效应；针对现有研究中协同效应测量或估计较少考虑团队成员历史合作情况的问题，本文充分利用球员之间的合作历史数据，提出平均值法、最近邻法和构成法三种算法对潜在团队的协同效应进行估计；针对现有团队绩效/比赛胜负预测研究对团队构建指导不足的问题，本文提出了考虑协同效应的比赛胜负预测模型。基于上述研究内容，本文利用NBA真实比赛数据进行实验。发现在潜在团队的协同效应估计方面，构成法由于考虑信息较为全面，能够较为准确地完成协同效应的估计；考虑协同效应的比赛胜负预测模型实现了利用历史赛季信息预测新赛季比赛的胜负，预测准确率达到67.92%，同现有利用上半赛季信息预测下半赛季比赛胜负的研究结果基本可比，同时实现了为球队管理层提供阵容团队构建的决策支持。相较现有研究，本研究提出了考虑成员之间合作历史情况，量化分析团队成员协同效应的可行性方案；将协同效应引入传统团队绩效预测以及团队成员选择问题的框架中，并证明了协同效应引入的有效性。本研究的成果可直接应用至篮球队团队管理方面，亦可为扩展至企业团队管理领域，为团队管理者提供决策支持。
Due to the benefits of synergistic effect, the team, as a form of organization, has been widely used in the enterprise environment. Under the circumstances of big data age, the use of data analytic tools to form efficient team, to achieve synergistic effect, to maximize team performance has become an important research topic. However, due to the difficulty of measurement on synergistic effect, most of previous research only consider information of individual team member in their frameworks. The synergistic effect between team members is often not included in the overall research framework and considered only as an explanatory factor between independent variables and team performance. Moreover, when evaluating synergistic effect in previous studies, researchers mainly consider the individual personality traits of team members. Only few researchers would consider the historical cooperation data between team members.Based on data availability, academic and industrial background, we choose National Basketball Association (NBA) as our research context, use data analytic tools to quantify synergistic effect between team members based on their cooperation history, and to predict team performance considering synergistic effect. We believe our study will help guide the team formation. In this paper, we put forward the measurement of the synergistic effect between basketball lineup members; We make full use of the historical cooperation data between players and put forward three algorithms (Average、Nearest Neighbor & Combination) to evaluate the synergistic effect between basketball lineup members; We also put forward a basketball team performance forecasting framework considering the synergistic effect. To evaluate our research framework, we use real NBA games as our experiment data set and find out that when evaluating synergistic effect of potential team members, Combination Algorithm has the best performance. When predicting NBA games considering synergistic effect, we successfully predict games using information from past seasons and achieve comparable forecasting performance. The predicting accuracy reaches up to 67.92%. We also successfully guide the team formation using one real example.In conclusion, compared with previous research, we put forward feasible schemes to quantify the synergistic effect between team members considering historical cooperation. We also introduce the synergistic effect into traditional research framework on team performance prediction and team member selection, and prove the effectiveness of the introduction of the synergistic effect. The results of this study can be directly applied to the basketball team management and extended to the enterprise team management field, providing decision support for team managers.
