HOW TO CHANGE THE STATEMENT OF A REGRESSION TASK A LITTLE AND GET A DEMING REGRESSION, OR HOW TO ESTIMATE A CLOUD OF POINTS
In this article, I approximate the points by a straight line under the assumption that there are no errors. Then I compare this result with the solution of the Deming Regression task.
To begin with, we defined a task to solve. The task was to find a line that describes a cloud of points in the assumption, that there are no errors. Then we concocted out the relations between b and k and derived an equation for k. The roots of this equation are critical points, and at some points, the minimum is achieved. Also, how I mentioned, that this equation always has a positive root.
At the second, we took a second to compare results with the solution of the Deming Regression task and found out that relations between k and b are the same. However, a simple experiment revealed that two straight lines, gained by two tasks, are different.
At the third, the work hasn’t ended yet. Further researches should be with more rigorous details that I missed. Also, there should be more statistically significant conclusions about the differences between the two solutions.
- [Jen07] A. C. Jensen – Deming regression, 2007.
- [Kan21] I. Y. Kaniovska – Synopsis of the probability theory and mathematical statistics lectures, 2021.