DAST-SR:群智感知中具有地理相关性的连续长尾数据的精确聚合方法

DAST-SR:An Accurate Data Aggregation Mechanism through Exploiting Sequential Long-tail and Spatial Relativity in Crowdsensing

大数据时代,众包系统需要通过聚合多个数据提供者的数据来获得准确的真相。在基于指纹识别的蓝牙定位应用场景中,通过对数据的长尾特性、连续性和地理关系的研究,提出了一种处理具有地理相关性的连续长尾数据的机制,即an accurate data aggregation mechanism processing sequential long-tail data with spatial relativity(DAST-SR)。为了捕获数据的长尾特性,该机制使用数据源出现错误的置信上限来估计可信度。而为了捕获数据的连续特性和地理相关性,该机制联合使用数据源提供的数据、前一时刻的聚合真相和相关实体的聚合真相作为虚拟源,聚合获得真相。通过虚拟数据集上的仿真,与an accurate data aggregation mechanism incorporating sequential long-tail characteristics(DAST)、Dynamic Truth Discovery(DynaTD)和truth discovery on correlated entities(TD-corr)相比,DAST-SR聚合结果的平均绝对误差和均方根误差最小,聚合的结果更加准确。

In the era ofbig data,information about the same objects can be gathered and accumulated from multiple sources(i.e.,crowdworkers)through so-called crowdsensing.Especially,in the indoor positioning system using Bluetooth fingerprints,dealing with the sequential long-tail data with spatial correlation,an accurate data aggregation mechanism processing sequential long-tail data with spatial relativity(DAST-SR)is put forward.The estimated confidence interval is used to infer each source's credibility.Meanwhile,the previous accumulated data,as well as the accumulated data for spatially correlated objects,are used as virtual sources to obtain the current aggregated value.Simulations using artificial data demonstrate that compared with an accurate data aggregation mechanism incorporating sequential long-tail characteristics(DAST),Dynamic Truth Discovery(DynaTD)and truth discovery on correlated entities(TD-corr),the mean absolute error(MAE)and the root mean square error(RMSE)of DAST-SR are both the smallest,which means it have the best performance.