The big data market solves the problem of the effective utilization of data through treating data as the circulating commodity in the market. The existing body of research on the big data market suggests that either improving the availability of published data or protecting sensitive information when trading data s the current mainstream topic. To date, the balancing the availability and privacy of the released dataset is a gradually emerging challenge. Unfortunately, there are few studies that have concentrated on the combination of the two points, which is more in line with the actual trading demands and data interaction patterns. Our paper proposes a novel mechanism called Differential Privacy Data Trading (DPDT) mechanism by introducing the differential privacy into the data trading process. Our DPDT mechanism can meet the actual usage requirements of data consumers for the released dataset while ensuring privacy. In short, the DPDT mechanism balances availability and privacy by generating a private synthetic dataset whose accuracy is determined by the data consumer. It is customized for the big data market by improving appropriate synthetic dataset privacy releasing techniques. In addition, DPDT can calculate the corresponding security payment costs depending on the different accuracy of the released dataset by correlating the accuracy, privacy budget, and payment. We instantiate the DPDT with real-world data and the experimental results verify the proposed mechanism is feasible and robust. Our analysis and discussion results reveal that DPDT achieves the trade-off between availability and privacy of the released dataset during data trading.

大数据市场通过将数据视为市场中的流通商品来解决数据有效利用的问题。关于大数据市场的现有研究表明，要么提高已发布数据的可用性，要么在交易数据时保护敏感信息是当前的主流话题。迄今为止，平衡已发布数据集的可用性和隐私性是一个逐渐出现的挑战。不幸的是，很少有研究集中在这两点的结合上，这更符合实际的交易需求和数据交互模式。通过将差异隐私引入到数据交易过程中，我们提出了一种新颖的机制，称为差异隐私数据交易（DPDT）机制。我们的DPDT机制可以在确保隐私的同时满足已发布数据集的数据使用者的实际使用需求。简而言之，DPDT机制通过生成私有合成数据集来平衡可用性和隐私，该合成数据集的准确性由数据使用者确定。通过改进适当的综合数据集隐私发布技术，为大数据市场量身定制了它。此外，DPDT可以通过将准确性，隐私预算和付款相关联，根据已发布数据集的不同准确性来计算相应的安全付款成本。我们使用实际数据实例化了DPDT，实验结果证明了该机制的可行性和鲁棒性。