Estimating regression models with unknown break-points

Area of Science:

• Statistics
• Econometrics
• Data Science

Background:

• Regression models are widely used for data analysis.
• Identifying structural changes or breakpoints is crucial in many applications.
• Existing methods for breakpoint analysis can be complex or computationally intensive.

Purpose of the Study:

• To develop a straightforward linearization technique for fitting piecewise regression models.
• To estimate models where breakpoints are true parameters.
• To extend the applicability to threshold modeling with explanatory variables.

Main Methods:

• A simple linearization technique is employed for estimation.
• The method leverages the linear formulation of piecewise regression.
• It is compatible with existing statistical software for linear models.

Main Results:

• The proposed method provides a computationally efficient way to fit piecewise regression models.
• It allows for the direct estimation of breakpoint parameters.
• The technique is demonstrated to be suitable for various regression models with linear predictors.

Conclusions:

• The linearization technique offers a practical and accessible approach to breakpoint analysis in regression.
• It simplifies the estimation of piecewise models, including threshold models.
• The method's compatibility with current software enhances its utility for researchers and practitioners.