Title :
Processing of Missing Values Using Gibbs Sampling
Author :
Yi, Wang ; Li, Zhou
Author_Institution :
Glorious Sun Sch. of Bus. & Manage., Donghua Univ. (DHU), Shanghai, China
Abstract :
In data warehousing Gibbs Sampling is applied in missing value processing due to the many defects in the traditional methods. As long as Full Conditional Distribution Condition (FCDC) is met, Gibbs sampling can solve issues such as high workload and biased data. The method´s high operability - it even can be completed in common used tool like Excel - makes it a practical method for real data preprocessing.
Keywords :
Markov processes; Monte Carlo methods; data handling; data warehouses; sampling methods; Excel; FCDC; Gibbs sampling; Markov Chain pattern; Monte Carlo-based approach; data warehousing; full conditional distribution condition; missing value processing; real data preprocessing; Analytical models; Bayesian methods; Business; Computational modeling; Data mining; Data models; Markov processes; Data warehouse; Full Conditional Distributions Condition; Gibbs sampler; Missing value processing;
Conference_Titel :
Measuring Technology and Mechatronics Automation (ICMTMA), 2011 Third International Conference on
Conference_Location :
Shangshai
Print_ISBN :
978-1-4244-9010-3
DOI :
10.1109/ICMTMA.2011.514
