An on-line algorithm for fitting straight lines between data ranges

ABSTRACT

Applications often require fitting straight lines to data that is input incrementally. The case where a data range [&agr;k, &ohgr;k] is received at each tk, t1 < t2 < … tn, is considered. An algorithm is presented that finds all the straight lines u = mt + b that pierce each data range, i.e., all pairs (m, b) such that &agr;k ≤ mtk + b ≤ &ohgr;k for k = 1, … , n. It may be that no single line fits all the ranges, and different alternatives for handling this possibility are considered. The algorithm is on-line, producing the correct partial result after processing the first k ranges for all k < n. For each k, the set of (m, b) pairs constitutes a convex polygon in the m-b parameter space, which can be constructed as the intersection of 2k half-planes. It is shown that the O(n logn) half-plane intersection algorithm of Shamos and Hoey can be improved in this special case to O(n).

REFERENCES

• 1: Freedman, A.M., Buneman, O.P., Peckham, G., and Trattner, A. Automatic recognition of significant events in the vital signs of neonatal infants. Comput. Biomed. Res. 12, 2 (Apr. 1979), 141-148.
• 2: D. T. Lee , F. P. Preparata, An Optimal Algorithm for Finding the Kernel of a Polygon, Journal of the ACM (JACM), v.26 n.3, p.415-421, july 1979 [doi>10.1145/322139.322142]
• 3: O'Rourke, J., and Badler, N. Model-based image analysis of human motion using constraint propagation. IEEE Trans. on Pattern Analysis and Machine Intelligence, PAM1-2, 6 (Nov. 1980), 522-536.
• 4: Pavlidis, T. Structural Pattern Recognition. Springer-Verlag, Berlin, 1977, pp. 168-184.
• 5: F. P. Preparata, An optimal real-time algorithm for planar convex hulls, Communications of the ACM, v.22 n.7, p.402-405, July 1979 [doi>10.1145/359131.359132]
• 6: Shamos, M.I., and Hoey, D. Geometric intersection problems. 17th Ann. IEEE Symp. Foundations of Computer Science, (Oct. 1976), 208-215.
• 7: Michael Ian Shamos, Computational geometry., 1978

INDEX TERMS

Primary Classification:
  G. Mathematics of Computing

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE

General Terms:
Algorithms, Design, Management

Keywords:
computational complexity, computational geometry, half-plane intersection, incremental algorithms, on-line algorithms, piecewise linear approximation