Intrinsic Dimensionality Estimation for Data Points in Local Region
- 21 Downloads
Intrinsic dimensionality estimation plays a pivotal role in dealing with high-dimensional datasets. In this work, we aim to develop a robust dimensionality estimation algorithm by investigating the intrinsic dimensionality estimation methods for data points in its local region. Our method is able to effectively utilise the geometric information in the local region for dimensionality. We also show different methods to improve the estimation by using perspectives from the local region and different preprocessing methods.
Keywords and phrasesNumerical analysis Probabilistic methods Models of computation Nonparametric inference Estimation
AMS (2000) subject classificationPrimary 62G05 Secondary 62P30
Unable to display preview. Download preview PDF.
This work is supported by the Natural Science Foundation of the Jiangsu Higher Education Institutions of China, No.15KJD110001
- Ceruti, C., Bassis, S., Rozza, A., Lombardi, G., Casiraghi, E. and Campadelli, P. (2012). DANCo: Dimensionality from angle and norm concentration. arXiv:1206.3881.
- Costa, J.A. and Hero, A.O. (2005). Learning intrinsic dimension and entropy of shapes. In Krim, H. and Yezzi, T. (eds.), Statistics and analysis of shapes. Basel: Birkhuser.Google Scholar
- Donoho, D.L. (2000). Aide-memoire. High-dimensional data analysis: the curses and blessings of dimensionality. American Math Society Lecture-Math Challenges of the 21st Century.Google Scholar
- Farahmand, A.M. and Szepesvri, C. (2007). Manifold-adaptive dimension estimation. In ICML 07: Proceedings of the 24th international conference on Machine learning, pp. 265–272.Google Scholar
- Hein, M. (2005). Intrinsic dimensionality estimation of submanifolds in euclidean space. In Proc. of ICML, pp. 289–296.Google Scholar
- Hubner, U., Weiss, C.O., Abraham, N.B. and Tang, D. (1994). Lorenz-like chaos in NH 3-FIR lasers, Time Series Prediction. Forecasting the Future and Understanding the Past, 73-104. Addison Wesley.Google Scholar
- Levina, E. and Bickel, P.J. (2005). Maximum likelihood estimation of intrinsic dimension. Adv. Neural Inf. Process. Syst.17, 777–784.Google Scholar
- Yu, K., Zhang, T. and Gong, Y. (2009). Nonlinear learning using local coordinate coding. Adv. Neural Inf. Process. Syst. Google Scholar