Saliency Score-Based Visualization for Data Quality Evaluation

다국어 초록 (Multilingual Abstract)

Data analysts explore collections of data to search for valuable information using various techniques and tricks. Garbage in, garbage out is a well-recognized idiom that emphasizes the importance of the quality of data in data analysis. It is therefore crucial to validate the data quality in the early stage of data analysis, and an effective method of evaluating the quality of data is hence required. In this paper, a method to visually characterize the quality of data using the notion of a saliency score is introduced. The saliency score is a measure comprising five indexes that captures certain aspects of data quality. Some experiment results are presented to show the applicability of proposed method.

참고문헌 (Reference)

1 G. Merz, "UCI repository of machine learning databases" Department of Information and Computer Science, University of California 1996

2 A. D. Chapman, "Principles of Data Quality" Global Biodiversity Information Facility 2005

3 C. M. Bishop, "Pattern Recognition and Machine Learning" Springer 2006

4 H. P. Kriegel, "Outlier detection techniques" 2009

5 K. P. Murphy, "Machine Learning: A Probabilistic Perspective" MIT press 2012

6 M. M. Breunig, "LOF: identifying density-based local outliers" 93-104, 2000

7 D. M. Hawkins, "Identification of Outliers" Chapman and Hall 1980

8 B. Heinrich, "How to measure data quality? A metric-based approach" 1-15, 2007

9 E. M. Knorr, "Finding intensional knowledge of distance-based outliers" 211-222, 1999

10 US Environmental Protection Agency, "Data quality assessment:statistical methods for practitioners" US Environmental Protection Agency 2006

11 Leo L. Pipino, "Data quality assessment" Association for Computing Machinery (ACM) 45 (45): 211-, 2002

12 C. C. Aggarwal, "Data Mining: The Textbook" Springer 2015

13 R. Y. Wang, "Beyond accuracy : what data quality means to data consumers" 12 (12): 5-33, 1996