| Budget Amount *help |
¥16,120,000 (Direct Cost: ¥12,400,000、Indirect Cost: ¥3,720,000)
Fiscal Year 2017: ¥5,330,000 (Direct Cost: ¥4,100,000、Indirect Cost: ¥1,230,000)
Fiscal Year 2016: ¥5,330,000 (Direct Cost: ¥4,100,000、Indirect Cost: ¥1,230,000)
Fiscal Year 2015: ¥5,460,000 (Direct Cost: ¥4,200,000、Indirect Cost: ¥1,260,000)
|
|---|
| Outline of Final Research Achievements |
In the era of Big Data, data volumes have become so enormous and so complex as to preclude processing using traditional applications. For similarity search and retrieval, as well as many other fundamental operations in such areas as data mining, machine learning, multimedia, recommendation systems, and bioinformatics, the efficiency and effectiveness of software implementations depends crucially on the interplay between measures of data similarity and the features (or attributes) by which data objects are represented. When the number of features (the data dimensionality) is high, the errors introduced into similarity measurements by the many irrelevant feature attributes can overwhelm the contributions of the relevant features. The overall goal of this project is to tackle the problem of the curse of dimensionality in similarity applications for big data, by developing practical unsupervised techniques that recognize and take advantage of local variations in intrinsic dimensionality.
|
