| Title | A Data Reconstruction Method for The Big-Data Analysis |
| Publication Type | Conference Paper |
| Year of Publication | 2018 |
| Authors | Mito, M., Murata, K., Eguchi, D., Mori, Y., Toyonaga, M. |
| Conference Name | 2018 9th International Conference on Awareness Science and Technology (iCAST) |
| Keywords | ART, Big Data, Big Data analysis tools, Big Data approach, big data privacy, big-data, big-data reconstruction method, correlation coefficient, Correlation-coefficient, Data analysis, data privacy, extreme value record elimination, Histogram, Histograms, privacy, privacy data, privacy problems, privacy-preserving methods, pubcrawl, Reconstruction algorithms, self organizing map, self-organising feature maps, SOM, SOM (Self-Organizing-Map) |
| Abstract | In recent years, the big-data approach has become important within various business operations and sales judgment tactics. Contrarily, numerous privacy problems limit the progress of their analysis technologies. To mitigate such problems, this paper proposes several privacy-preserving methods, i.e., anonymization, extreme value record elimination, fully encrypted analysis, and so on. However, privacy-cracking fears still remain that prevent the open use of big-data by other, external organizations. We propose a big-data reconstruction method that does not intrinsically use privacy data. The method uses only the statistical features of big-data, i.e., its attribute histograms and their correlation coefficients. To verify whether valuable information can be extracted using this method, we evaluate the data by using Self Organizing Map (SOM) as one of the big-data analysis tools. The results show that the same pieces of information are extracted from our data and the big-data. |
| DOI | 10.1109/ICAwST.2018.8517197 |
| Citation Key | mito_data_2018 |
A Data Reconstruction Method for The Big-Data Analysis