A RANDOMIZED ALGORITHM FOR SLOPE SELECTION

被引:37
作者
Dillencourt, Michael B. [1 ]
Mount, David M. [2 ,3 ]
Netanyahu, Nathan S. [4 ]
机构
[1] Univ Calif Irvine, Dept Informat & Comp Sci, Irvine, CA 92717 USA
[2] Univ Maryland, Dept Comp Sci, College Pk, MD 20742 USA
[3] Univ Maryland, Inst Adv Comp Studies, College Pk, MD 20742 USA
[4] Univ Maryland, Ctr Automat Res, Comp Vis Lab, College Pk, MD 20742 USA
基金
美国国家科学基金会;
关键词
Slope selection; randomized algorithm; Theil-Sen estimator; line fitting; robust estimation;
D O I
10.1142/S0218195992000020
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
A set of n distinct points in the plane defines ((n)(2)) lines by joining each pair of distinct points. The median slope of these O(n(2)) lines was proposed by Theil as a robust estimator for the slope of the line of best fit for the points. We present a randomized algorithm for selecting the k-th smallest slope of such a set of lines which runs in expected O(n log n) time. An efficient implementation of the algorithm and practical experience with the algorithm are discussed.
引用
收藏
页码:1 / 27
页数:27
相关论文
共 31 条
[1]  
Aho A. V., 1974, DESIGN ANAL COMPUTER
[2]   FAST PROBABILISTIC ALGORITHMS FOR HAMILTONIAN CIRCUITS AND MATCHINGS [J].
ANGLUIN, D ;
VALIANT, LG .
JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 1979, 18 (02) :155-193
[3]  
[Anonymous], 1950, P R NETHERLANDS ACAD
[4]  
[Anonymous], 2003, ROBUST REGRESSION OU
[5]  
BENTLEY JL, 1990, PROCEEDINGS OF THE FIRST ANNUAL ACM-SIAM SYMPOSIUM ON DISCRETE ALGORITHMS, P91
[6]   AN OPTIMAL-TIME ALGORITHM FOR SLOPE SELECTION [J].
COLE, R ;
SALOWE, JS ;
STEIGER, WL ;
SZEMEREDI, E .
SIAM JOURNAL ON COMPUTING, 1989, 18 (04) :792-810
[7]  
Dahlquist G., 1974, NUMERICAL METHODS
[8]  
Donoho D. L., 1983, FESTSCHRIFT L LEHMAN, P157
[9]   USE OF HOUGH TRANSFORMATION TO DETECT LINES AND CURVES IN PICTURES [J].
DUDA, RO ;
HART, PE .
COMMUNICATIONS OF THE ACM, 1972, 15 (01) :11-&
[10]   COMPUTING LEAST MEDIAN OF SQUARES REGRESSION LINES AND GUIDED TOPOLOGICAL SWEEP [J].
EDELSBRUNNER, H ;
SOUVAINE, DL .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1990, 85 (409) :115-119