In this paper, we consider fitting a line segment to noisy data. We describe the structural segment model and establish conditions for its identifiability. The method of moments estimator (MME) is introduced and explored. The MME is easily computed and is invariant under translation rotation and reflection. We show that the MME follows, asymptotically, a normal distribution. The asymptotic efficiency of the MME relative to the maximum likelihood is investigated numerically and found to be high in most cases.
- Line-segment structural relationship
- Measurement error
- Method of moments
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty
- Applied Mathematics