大数据的应用核心就是大数据预测。大数据预测完全依赖大数据来源，因此具有“全样非抽样、效率非精确、相关非因果”的特征。按照预测的精细程度，大数据预测的层级分为三级，依次递进：圈子范围的大小、圈子内有哪些族群、族群中每个成员的个性特征。能否在三个层次获得准确的预测结果，关键在于前台数据和后台数据、宏观数据和微观数据、共性数据和个性之间的关联分析。对于任何大数据的预测结果，必须设计一个验证流程，以防止出现不必要的错误。

The core application of big data is forecasts based on big data. It is accuracy entirely dependent on big data sources, and therefore has the feature of "full but non-part, efficiency but non-precise, relevant but non-causal". According to forecasts fineness, big data predicted l accuracy is divided into three levels, followed by progressive: the size of the groups,the size of sub- groups, the personality characteristics of each member within the sub-groups. The key to get the accurate predictions is correlation analysis between the behavioral data and outcome data, the macro data and micro data, and the common and personality data. For any predictions of big data, you must design a verification process to prevent unnecessary errors.