[ArXiv] 3rd week, Apr. 2008

hlee — Mon, 21 Apr 2008 01:05:55 +0000

The dichotomy of outliers; detecting outliers to be discarded or to be investigated; statistics that is robust enough not to be influenced by outliers or sensitive enough to alert the anomaly in the data distribution. Although not related, one paper about outliers made me to dwell on what outliers are. This week topics are diverse.

[astro-ph:0804.1809] H. Khiabanian, I.P. Dell’Antonio
A Multi-Resolution Weak Lensing Mass Reconstruction Method (Maximum likelihood approach; my naive eyes sensed a certain degree of relationship to the GREAT08 CHALLENGE)
[astro-ph:0804.1909] A. Leccardi and S. Molendi
Radial temperature profiles for a large sample of galaxy clusters observed with XMM-Newton
[astro-ph:0804.1964] C. Young & P. Gallagher
Multiscale Edge Detection in the Corona
[astro-ph:0804.2387] C. Destri, H. J. de Vega, N. G. Sanchez
The CMB Quadrupole depression produced by early fast-roll inflation: MCMC analysis of WMAP and SDSS data
[astro-ph:0804.2437] P. Bielewicz, A. Riazuelo
The study of topology of the universe using multipole vectors
[astro-ph:0804.2494] S. Bhattacharya, A. Kosowsky
Systematic Errors in Sunyaev-Zeldovich Surveys of Galaxy Cluster Velocities
[astro-ph:0804.2631] M. J. Mortonson, W. Hu
Reionization constraints from five-year WMAP data
[astro-ph:0804.2645] R. Stompor et al.
Maximum Likelihood algorithm for parametric component separation in CMB experiments (separate section for calibration errors)
[astro-ph:0804.2671] Peeples, Pogge, and Stanek
Outliers from the Mass–Metallicity Relation I: A Sample of Metal-Rich Dwarf Galaxies from SDSS
[astro-ph:0804.2716] H. Moradi, P.S. Cally
Time-Distance Modelling In A Simulated Sunspot Atmosphere (discusses systematic uncertainty)
[astro-ph:0804.2761] S. Iguchi, T. Okuda
The FFX Correlator
[astro-ph:0804.2742] M Bazarghan
Automated Classification of ELODIE Stellar Spectral Library Using Probabilistic Artificial Neural Networks
[astro-ph:0804.2827]S.H. Suyu et al.
Dissecting the Gravitational Lens B1608+656: Lens Potential Reconstruction (Bayesian)

[ArXiv] Decision Tree, Aug. 31, 2007

hlee — Wed, 05 Sep 2007 02:55:57 +0000

From arxiv/astro-ph:0708.4274v1
Comparison of decision tree methods for finding active objects by Y. Zhao and Y. Zhang

The authors (astronomers) introduced and summarized various decision three methods (REPTree, Random Tree, Decision Stump, Random Forest, J48, NBTree, and AdTree) to the astronomical community.

The goal of applying decision tree methods is discriminating active objects (quasars, BL Lac objects, and active galaxies) from non-active objects (stars and galaxies) and overcoming drawbacks of popular neural networks (NNs) and support vector machine (SVM) in the astronomical society thanks to the following properties: non-parametric modeling (does not require strong model assumptions), identifiability of important independent variables, and relatively short training period suitable for huge data sets. Shortcomings of decision tree methods were also described. Nonetheless, the fact that the decision tree provides clear and easy interpretable classification rules is hardly ignorable.

Performances of the listed methods were compared based on their accuracy and computing time. For separating AGNs from non-AGNs, AdTree performed better in terms of accuracy and Decision Stump produced the fastest result. The software for this study was WEKA (The Waikato Environment for Knowledge Analysis), which is available from this link.

The AstroStat Slog » neural network

[ArXiv] 3rd week, Apr. 2008

[ArXiv] Decision Tree, Aug. 31, 2007